PUZZLE CORNER
ALLAN J. GOTTLIEB
Midsummer Snowstorm
Our group is just now moving to
a newly renovated floor. We had
a major influence in specifying
the layout, and it is satisfying to see how
well it all works. I guess that if our re
search efforts into computer architecture
turn out to fizzle we have a related field
to move into.
Problems
A/S 1. Here is a chess problem from The
Tech, M.I.T.'s student newspaper.
White is to move and force a mate in
two:
A/S 2. Another geometry problem from
Phelps Meaker, who writes:
I wanted an approximation of a four-foot
sphere for light-integration measure
ment. I made a dodecahedron of sheet
metal, except that instead of flat sur
faces, I substituted low five-sided pyr
amids. It served my purpose. How
should I have designed the pyramids so
that all the dihedral angles were equal?
A/S 3. Randy Barron poses a question
that should appeal to all EE majors:
SEND PROBLEMS. SOLUTIONS,
AND COMMENTS TO ALLAN /.
GOTTLIEB. '67, ASSOCIATE
RESEARCH PROFESSOR AT
THE COURANT INSTITUTE OF
MATHEMATICAL SCIENCES.
NEW YORK UNIVERSITY. 25]
MERCER ST.. NEW YORK.
NY.. 10012.
A four-port device contains only passive
linear circuit elements. Using a fixed-fre
quency sine-wave generator, you can
measure the complex impedance Z be
tween two of the terminals, A and B. A
variable resistor R is connected across
the other two terminals, C and D. What
is the locus in the complex plane traced
out by Z as R varies from zero (short
circuit) to infinity (open circuit)?
A/S 4. Here is a temporal cryptarithmetic
problem from Nob Yoshigahara relayed
to me via Richard Hess (on DICKNET in-
terpuzzler communication network):
Fill in the boxes with the digits 1,2,3.. .,9:
DD^OD" x □ = □□""": DD*"
A/S 5. An out-of-season problem from
Bruce Calder:
It began to snow on a certain morning,
and the snow continued to fall steadily
throughout the day. At noon, a snow-
plow started to clear a road at a constant
rate in terms of the volume of snow re
moved per hour. The snowplow cleared
two miles by 2 p.m. and one more mile
by 4 p.m. At what time had the snow
storm begun?
Speed Department
SD 1. Greg Huber sent us the following
problem, told to him by Douglas Hof-
stadter:
Simplify the product
(x - a)(x - b) . . . (x - z).
SD2. Doug Van Patter offers a bridge
quickie:
North: East:
AK93 A10 76
¥653 ¥ AKQ7
♦ AJ92 ♦ Q75
*Q74 *J109
You are East in a rubber bridge game;
bidding has gone: South INT, North
3NT. Your partner leads the ¥10, and
everyone follows to three rounds of
hearts. What is your best shot at setting
this game?
Solutions
APR 1. White is lo play and mate in two:
Jerry Grossman sent us the following solution:
White moves g5-f6 (en passant), then f6-f7 (check
mate) after any reply by Black. To prove that this
solution is correct, we first claim that Black's pre
vious move was either d7-d5 or (7-f5. These are
clearly the only pawn moves possible (since the
White king could not have been in check when Black
moved), but why couldn't the Black king have
moved to its present position on the last move to
escape check? It couldn't have come from d8 or f8,
since there would have been no way for White to
administer the double check (pawn at e7 and
knight). On the other hand, if the Black king came
from d7 or f7, Ihen White would have had to have
moved the pawn to e6 on the previous move but
there is no now-vacant square from which this
pawn could have come. Thus we have verified this
first claim. Now, in order for White's pawns to be
in the position they are in, they must have made at
least 10 captures. Black is missing only 10 pieces,
however, and hence every one must have been cap
tured by a pawn. If Black's last move was d7-d5,
then Black's queen bishop could not have ever
moved and could not have been captured by a
pawn. This contradiction proves that Black's last
move could only have been f7-f5, and thus the so-
AJO AUGUST/SEPTEMBER 1985
Clyde B. Dooiitlle, '23; Februar>- 22,1985; Hillsdale,N.J.
Thomas B. Drew, '23; May 5, 1985; Peterborouch,N.H. hWilliam J. O'Shaughnessey, '23; July 22, 1984; Ma-con, Ga.
Clinton B. Conway, '24; April 11, 1985; Altoona.Fla.
Robert Bruce Lindsay, '24; February 2, 1985; Providence, R.I.
Mrs. Sidney W. Andrews, '25; January 25, 1985;South Salem, N.Y.
Alexander C. Brown, '25; January 13. 1985; Cincinnati, Ohio.
George L. Bums, '25; 1984; Portland, Ore.
Frederic S. Jacques, '25; March 22, 1985; Bangor,Maine.
Augustine J. Cotter, '26; 1984; Scarsport, Maine.Valentine F. Harrington, '26; 1985.
Richard L. O'Donovan, '27; April 23, 1985; CoralGables, Ha.
Helmuth G.R. Schneider, '27; February 25, 1985-Westfield, N.J.
Darcy A. Young, Jr., "27; November 24, 1984; Peterborough, N.H.
Alexander B. Daytz, '28; December 1984; Chats-worth, Calif.
Mrs. David Mathoff, '28; 1985; Boston, Mass.Harold G. Nyman, '28; June 2,1984; Belmont, Mass.
Waller R. Ramsaur, '28; May 11, 1984; Pacific Palisades, Calif.
Joseph D. Riley, '28; March 19, 1985; LighthousePt., Fla.
Francis C. Sweeney, '28; March 15, 1985; Clifton.N.J.
Charles B. Bacon, '29; April 30, 1985; Middletown,Conn.
William E. Lowery, '29; November 29, 1984; Plymouth, Mass.
Henry S. Muller, '29; March 26, 1985; Belmont,Ohio.
Henry N. Bates, '30; January 1985; Los Gatos, Calif.
Thomas J. Hickey, '30; January 18, 1985; Arlington,Va.
Jonathan G. Swift, "30; January 18,1985; West Hartford, Conn.
Albert Earl Cullum, Jr., '31; January 31, 1985; Dallas, Tex.
Joseph H. Passell, '31; March 15,1985; Williamston,
N.CIra J. Bach, '32; March 6, 1985; Chicago, 111.
EdwinC. Beck, '32; May 31,1984; Muskegon, Mich.
Harry F. Carlson, '32; August 31, 1984; Hanover,Mass.
Maurice D. Triouleyre, '32; April 26, 1985; Long-meadow, Mass.
Joseph Welch, Jr., '32; April 13. 1985; Fort Lauder-dale, Fla.
Weston L. Brannen, '33; March 14,1978; Valparaiso,Ind.
J. Mason Culverwell, '33; November 3, 1984; Wash
ington, D.C.
Leo V. Dewar, '33; November 5. 1984; Rochester,
N.Y.
John R. Hopkins, '33; March 16, 1985; Los Alamos,N.M.
Kenneth D. Moslander, '33; February 1985; FortMyers, Fla.
Robert N. Eck, '34; March 29, 1985; Brookfield.Wise.
Waller R. Hedeman, Jr., '34; April 14, 1985; An
napolis, Md.
Octavio Leonard Colavecchio, '35; April 2, 1985;Providence, R.I.
Dwighl P. Merrill, '35; December 27,1984; Newton,
Mass.
Joseph 1. Ackerman, Jr., '36; May 9, 1985; Arling
ton, Mass.
Wilber W. Haynes, '38; 1985; Kissimmee, Fla.
Alfred Sweeney, Jr., '38; December 13, 1984; Santa
Barbara, Calif.
Robert G. Larkln, '39; February 15, 1985; VenturaCalif.
Ramon S. Sevilla, Sr., '39; April I, 1985; Metro Manila, Philippines.
Ernest G. Chilton, '40; January 22, 1985; MenloPark, Calif.
David Joseph Collins, '40; April 13,1985; Wcllesley,Mass.
J. Herbert Hollomon, '40; May 8, 1985; Brookline,Mass.
Spencer M. Richardson, '40; May 26. 1984; Quak-erlown, Penn.
Richmond W. Wilson, '40; February 7, 1985; Corning, N.Y.
Richard H. Gould, Jr., '41; 1982; Douglaston, N.Y.
Floyd W. Iden, '41; June 27, 1984; Pompton Plains,N.J.
Robert P. Boyer, '42; December 4, 1984; Manchester, Tenn.
Joseph A. Crutcher, '42; April 1985; Fullerton, Calif
Wesley R. Floyd, '42; February 29, 1984; Bradenton,Fla.
Ray O. Wyland, Jr., '42; April 1, 1985; Tujunga,Calif.
Margaret L. Blizard, '43; May 10, 1985; Norwood.Mass.
Clyde A. Booker, Jr., '43; 1985; Pittsburgh, Penn.
Shao T. Hsu, '43; 1985; Bethesda, Md.
James B. Hoaglund, '45; February 9, 1985; Minne
apolis, Minn.
David A. Trageser. '45; April 26, 1985; Wayland.Mass.
William H. Auerswald, '46; April 10, 1985; Suffield,Conn.
Jacob W. Ullmann, '46; February 5.1985; New York,N.Y.
Helmar Schlein, '47; February 19, 1985; Reseda,
Calif.
John F. Christopher, '48; April 24, 1985; St. Louis,
Mo.
Ralph Segal, '48; January 1985; Roslyn, N.Y.
David S. Selengut, '48; July 24, 1984; Schenectady,N.Y.
Hcndrie J. Grant, '49; August 1, 1984; St. Paul,
Minn.
Leonard N. McKibben, '49; December 31, 1984;
Westlake, Ohio.
Waller E. Mutter, '49; November 1984; Poughkeep-sie, N.Y.
Mctr Drubin, '50; April 25, 1985; Rancho Palos
Verde, Calif.
Robert S. Lovett, '50; June 1984; Wilmington, Del.
Dean E. Cogswell, '51; May 11, 1985; Wenham,
Mass.
Bryant W. Foster, '51; March 17, 1985; Braintree,
Mass.
Charles G. Etler, Jr., '53; December 22, 1983;
Wayne, Penn.
Lawrence Tsun-Ying Kwan, '53; December 1967;
Encino, Calif.
Alan K. MacKenzie, '54; April 10, 1985.
Tse H. Chu, '55; March 1982.
Agnes M. Galligan, '55; 1985; West Roxbury, Mass.
Jakob Auslaender, '56; February 5, 1985; Chicago,
III.
Waller C. Kottemann, '59; Mav 5. 1984. Clarendon
Hills. 111.
Vinod Sundra, '59; December 20, 1980; Bedford,
Mass.
Michael Vincent Bogda, '66; May 22, 1983; Wood
land Hills, Calif.
Paul G. Maguire, 68; April 4. 1984; Aclon, Mass.
Theodore R. Sieger, Jr., '74; 1985; Longmeadow.
Mass
Haley&Aldnch,Inc
Consulting Oootochnlcal
Engineers and
Geologists
Soil and Rock
Mechanics
Engineering Geology
Engineering Goophystcs
Foundation Engineering
Terrain Evaluation
Engineering Seismology
Earthquako Engineering
Geohydrology
238 Main St.
Cambridge. MA 02142
(617) 492-6460
Hsrl P. Aldrlch, Jr. '47
Martin C. Murphy '51
Edward B. Klnrtar '67
Douglas G. Gilford '71
Joseph J. Rimer '68
John P. Dugan '68
Kenneth l_ Rocker 73
Mark X. Haley 75
Robin B. Dill '77
Andrew F. McKown 78
Keith E. Johnson '80
Hawkins&SansGampany
Buitoln£ contractors
Steven H. Hawkins, '57
188 Whiting Street
Hlngham, MA 02043
(617)749-6011
(617) 749-6012
LnterfechE
Associates,
Hardware/Softwaro
Research &
Development
• Real Time Micro
Processor Based
Designs
• Data Acquisition
and Signal
Processing
• Microcomputer
Applications
Systems
Oovld A. Vogel, Ph.D
74
Dan C. Vogel. SM '81
(617) 329-0300
• Communications
and Networking
• Distributed
Processor
Designs
« Simulations
Exparloncod wllh most
micros, minis,
languages and
operating systems.
Ames ProfessionalCenter
450 Washington Street
Suite 103
Dedltam. MA 02026
TECHNOLOGY REVIEW A29
lution announced above provides the only male in
Iwo.
Also solved by Jeffrey Mattox, John Bobbin, Mat
thew Fountain, and Neil Hochstedler.
APR 2. Note that each of the following is a different
way of evaluating the same equation:
16601.92 + 14374.08 =
11334.4 + 19641.6 =
18521.44 + 12454.56 =
4147.36 + 26828.64 =
Write the equation in the usual form.
John Bobbitt was able to recognize the hidden
quadratic and writes:
The four "sets of numbers" represent
a! + 2ab + b! = (a + b)1where a = 111.6 and b a 64.4. By forming different
groupings of the four terms a!, b2, ab, ab, we cangel the four sets. Thus,
(a2 + b!) + 2ab = (a + b)!(ab + b2) + (ab + a1) = (a + b)2
(b2 + 2ab ) + a2 «■ (a + b)2b2 + (a2 + 2ab ) = (a + b)2
Also solved by Harry Zaremba, Jerry Grossman,
Marshall Fritz, Matthew Fountain, and Winslow
Hartford.
APR 3. Find all maxima and minima of
f)In/1 + e" +
m(i + e- - I)
without using calculus.
The solution below is from George Bird:
Note that the expression (1 + e") may be written
as(l + 1/e"). Use the second form of this expression
and write it as a fraction with the common denom
inator e", thus:
(e1 + l)/e*
If we substitute this form, the expression
ln(I + e"*) becomes:
ln(l 4 e") - ln(e»).
The function
|ln(l + e"") + x/2|/|ln(l + e") - x/2]
then becomes:
|ln(e' + 1) - ln(e') + x/21 / [ln(l + el) - x/2|,
which reduces to 1/1. Therefore there are infinitely
many maxima and minima, and all have value 1.
Also solved by Allen Tracht, Edwin McMillan,
Harry Zaremba, Howard Stern, J. Richard Swen-
son. Jerry Grossman, John Bobbin, John Trussing,
Ken Haruta, Marshall Fritz, Matthew Fountain,
Mike Hennessey, Naomi Markovitz, Peter Card,
Ross Hoffman, Steve Feldman, Steve Silberberg,
Tony Trojanowski, Winslow Hartford, and the pro
poser. Rick Decker.
APR 4. On each day of the year (not leap year) you
are given a penny. On December 31 you are given
your last penny and told that it was fresh from the
U.S. Mint, but that one of the previous pennies may
have been counterfeit, and therefore lighter or heav
ier than the standard penny. You are asked to de
termine the number of balancings, using a common
pan balance, that would be necessary and sufficient
to determine whether or not there is a counterfeit
coin, and if there is, to tell whether it is heavier or
lighter than the last penny that you received.
The following solution is from Leon Tabak:
Two balancings are necessary (clearly) and suffi
cient for determining which of three possible so
lutions is the true situation:
(1) all coins are genuine.
(2) one coin is counterfeit and it is lighter that all of
the other (genuine) coins.
(3) one coin is counterfeit and it is heavier that all
of the other (genuine) coins.
Divide the set of 364 coins (whose genuineness is
not known) into three sets: SI, S2, and S3. Let SI
contain 122 coins. Let S2 and S3 each contain 121
coins. A fourth set, G, contains the one certified,
genuine coin.
First measurement: Place SI in the left pan of the
scale. Place S2 and G on the right side of the scale.
If the Iwo balance, then all of the coins in SI and
S2 must be genuine and the counterfeit coin, if there
is one, must be in S3. The possibilities that remain
Alexander
Kusko, Inc
Research, Aleiandar Kusko '44
development and
engineering services 161 Highland Avenue
In tho olcctrlcal Needhom Holghts, MA
engineering field 02194
(617)444-1381
Specialties:
Electric power
systems
Electric transportation
equlpmoni
Elsctrlc mochlnory
and magnetics
Solid-stats motor
drives, rectifiers.
Inverters
Feedback control
systems
Computer applications
and modollng
Evaluation,
Investigation,
patents.
LEA Group
LINENTHAL Consultants to
EISENBERQ Industry, commerce,
ANDERSON, INC. government and
ENGINEERS Institutions.
Building Oeslgn Eugerto R. Elsonberg
Environments!
Engineering *?»}• Ho«'.MISlte'Clvll Design Anderson 50
Roofing Technology S"llMnASpHa'1leXj52
75 Kneeland Street MSCEBoston, MA 02111 Vlesturs H. Ule '76
(617)426-6300
Suite 213
3811 Mallard Drive
Laurel, MD 20708
(301) 725-3445
LendElectric
Inc
Electrical contractors
to the nation since
1895
Headquarters:
45 Rockefeller Plaza
New York, N.Y. 10111
Offices In 16 principal
cities throughout the
U.S. and Puerto Rico
Boston Office:
66 Coolldgo Ave.
Watertown, MA 02172
(617)926-5500
Alexander W. Motfet. Jr.
Edward R
MaidenGup.
Builders for Industry,
Institutions, hospitals,
manufacturing plants,
QOVttffifTwnt flito
developers of high
technology facilities
for over 35 years
Edward R. Marden '41
Douglas R. Mardan '82
280 Lincoln Stroet
Boston, MA 02134
(617) 782-3743
Nelson,Coulson &
Associates,Inc
Professional Staffing Paul R. Coulson
Consultants PE, '43
President
Contract Engineering
Services
Professional
Recruiting Services
Technical Personnel In
All Disciplines
333 W. Hampden Ave.
Suite 507
Englowood, CO 80110
(303)761-7680
Other offices In
Albuquerque,
Colorado Springs,
DsHos & Seattle
Picragli&Associates
Corporate
Health Care
Giorgio PIccagli,PK fi *KTrit.U., Of
100 Dorado Terrace
San Francisco, CA
94112
(415) 333-5064
TECHNOLOGY REVIEW AJI
Geoage A.Roman &
Associates Ihc
Architecture Planning
Interior Design
Institutional
CofltimrcisI
Industrial
Residential
Slto Evaluation
Land lisa Planning
Master Planning
Programming
Interior Space
Planning
Colleges
Hospitals
Modlcal Buildings
Otflco Buildings
Apartments
Condominiums
George A. Roman,
A.LA. '65
Ono Gateway Cantor
Newton, MA 02158
(617) 332-5427
Robert F. RoweConstruction
Consultant
Robert F. Rowo '43
Suite 1-532
Control Data Business
and Technology
Center
Your building
program should
bo enjoyable.
II it's not
lei us smooth
out Ihe process.
35 years' 'experience in
Industrial,
commercial, and
institutional
construction.
701 East Bay SI.
Charleston. S.C. 29403
(803) 722-S610
StemhiedierCoip.
Contract research and
development in
radio frequency,
microwave and
mlltmeter wave
engineering and
related areas.
RF and Microwave
Systems Design
Industrial Applications
ol Microwave Power
Precision
Instrumentation
Analog and Digital
Electronics
Manufacturing
facilities available
185 New Boston Street
Wooum, MA 01601
Telex 948400
(617)935-8460
if the two pans balance in this first measurementare:
(1) all coins arc genuine.
(2) S3 contains an extra-light counterfeit coin.
(3) S3 contains an extra-heavy counterfeit coin.If the right pan comes to resl at a lower level than
the left pan, then two possibilities remain:
(1) SI contains an extra-light counterfeit coin.
(2) S2 contains an extra-heavy counterfeit coin.In either case, all of the coins in S3 must be genuine.
(There is at most one counterfeit coin, and this lesthas shown lhat it must be in SI or S2).
Second measurement: Place SI in Ihe left pan. Place
S3 and C in Ihe right pan If Ihe scale balances on
the second measurement and it also balanced on
the first measurement, then all coins are genuine.
If the left pan is higher, but the scale balanced dur
ing Ihe first measurement, then S3 must contain an
extra-heavy counterfeit coin. If the left pan is lower,
but the scale balanced during Ihe first measure
ment, Ihen S3 must contain an cxira-light counter
feit coin. If the scale balances on the second
measurement, but the left pan rested lower during
the first measurement, then S2 must contain an ex
tra-light counterfeit coin. If the scale balances on
the second measurement, but the left pan rested
higher during the first measurement, then S2 must
contain an extra-heavy counterfeit coin. If the left
pan is lower and the left pan was also lower during
the first measurement, then SI must contain an ex
tra-heavy counterfeit coin. If the left pan is higher
and the left pan was also higher during the first
measurement, then SI must contain an extra-light
counterfeit coin. Two outcomes are inconsistent
with Ihe statement of the problem. The left pan
cannot be higher in the first measurement and lower
M.I.T. ALUMNI
CAREER SERVICES
GazetteA listing every two weeks
of jobs for alumni across
the country
We want more firms to
know they can list jobs in
the Gazette
We want more alumni mak
ing a job change to know
the Gazette can direct them
to job opportunities
Whether you have a |ob to
fill, or are looking for a job,
let us send you a copy of
the Gazette to show you
how it can help you
Call or write
Marianne Ciarlo
Alumni Career Services
M.I.T., Room 12-170
Cambridge, Mass 02139
Tel: (617) 253-4735
in the second, nor can it be lower in Ihe first and
higher in the second. Either outcome would imply
the existence o( more than one counterfeit coin.
Also solved by Rik Anderson. Dudley Church,
E.P. Schacht, Frederic Jelen, Harry Zaremba, John
Prussing, John Spalding. Kenneth Olshansky. Mat
thew Fountain, Phelps Meaker, Walter Guett, and
the proposer, Alan Faller. Mr. Faller also showed
that with six balancings, one can also determine
which coin was counterfeit (except for leap years
when seven are required). Copies of Mr. Fa'ller'ssolution arc available upon request.
APR 5. Find X for Iwo configurations—when A ^
15, B = 10, and. H = 8; and when A " 16 + 2V2.
B = 16 - 2V2, and H = 2.
Howard Stem's solution is one of the few in
which an exact solution was found for the second
configuration:
With the above lengths labelled, similar triangles
give the following relationships:
(VB\ - X'VX = H/Y
(VA! - X2)'X = H/(X - Y).
Eliminating Y yields:
1/(V'B; - X;) - 1/(VAJ - X2) = 1/H (1)
The left side of (1), viewed as a function of X, has
a minimum of (l'A) + (!/B) when X = 0 and in
creases with X. For the first set of parameters given
(A = 15, B = 10, H = 8), the left hand side of (I)
will always be greater than 1/6 = 1/15 + 1/10. But
the right hand side is 1/8. Therefore, (1) can never
be satisfied; so these "crossed ladders" represent
an impossible configuration. The highest point of
crossing is 6, never 8.
The second set of parameters do allow for a phys
ically possible solution. After substituing A = 16
+ 2(2)'2, B = 16 - 2(2)" and H = 2 in (1), thesolution is X = (168)" = 12.96.
Also solved by Allen Tracht, Avi Ornslein, Fred
eric Jelen, Harry Zaremba, Matthew Fountain, Peter
Card, Ross Hoffman, Steve Feldman, Winslow
Hartford, and the proposer, Martin Brock.
Better Late Than Never
1984 N/D 1. Neil Hochstedler notes that the third
move m the third variation should be Q-a4 mate
1985 JAN 3. Randall Whitman has responded.
F/M 3, F/M 4. Ceorge Parks has responded.
APR SD2. E.P. Schact wants to rescind Horton's
Nobel Prize believing that the chicken came first.
Proposer's Soluiions to Speed Problems
SD 1. Zero.
SD 2. Don't cash your fourth heart! Lead a club and
trust that declarer now thinks that it is safe to take
the club finesse into your hand.
♦ K93
V 653
♦ AJ92
♦ Q74A J542
V 10 9 8
♦ 863
A 8653 A AQ8
VJ42
♦ K 10 4 3
A A K3
A 10 7 6
V A KQ7
♦ Q75
A J 10 9
Ml AUGUST/SEPTEMBER 1985
PUZZLE CORNER
ALLAN J. GOTTLIEB
Playing 20 Hands With
No Winners, and the Feeding of Cats
and Explorers
Since this isr the first issue of a new
academic year, I once more review
the ground rules under which this
department is conducted.
In each issue I present five regular
problems (the first of which is chess,
bridge, or computer-related) and two
"speed" problems. Readers are invited
to submit solutions to the regular prob
lems, and three issues later one sub
mitted solution is printed for each
problem; I also list other readers whose
solutions were successful. For example,
solutions to the problems you see below
will appear in the February/March issue.
Since I must submit that column some
time in November (today is July 19), you
should send your solutions to me during
the next few weeks. Late solutions, as
well as comments on published solu
tions, are acknowledged in the section
"Better Late Than Never" in subsequent
issues.
For "speed" problems the procedure
is quite different. Often whimsical,
these problems should not be taken too
seriously. If the proposer submits a so
lution with the problem, that solution
appears at the end of the same column
in which the problem is published. For
example, the solutions to this issue's
"speed" problems are given below.
Only rarely are comments on "speed"
problems published or acknowledged.
There is also an annual problem, pub
lished in the first issue of each new year;
and sometimes I go back into history to
republish problems that remained un
solved after their first appearance.
Problems
OCT1. We begin with a bridge problem
from Lawrence C. Kells, who writes that
a friend directed a duplicate bridge tour
nament in which every hand was played
20 times. One of these hands produced
a strange result. At four of the tables the
final contract was one club, played once
from each of the foursides. At four of
the other tables it was played at one dia
mond once from each side. At the re
maining tables it was played once from
each side at one heart, one spade, and
one no-trump. Every one of these con
tracts was set. Analysis of the hand
proved that none of the declarers made
a mistake in play. Unfortunately, Mr.
Kells failed to see what the deal was.
Can you reconstruct it? (That is, a deal
where any contract anybody bids can be
set no matter how hard he tries to make
it.)
OCT 2. Smith D. Turner (Jdt) has a
question about the following nim-like
game, called the game of thirty-one, that
was popular with New York City ma
gicians about 40 years ago:
Twenty-four cards, the 1-6 of each suit,
are put face-up on the table. Two players
pick up alternately, each keeping track
of his pip-total. One wins by hitting
thirty-one, or forcing one's opponent to
exceed 31. Does the first or second
player have a sure win, and how does
he play to insure it?
OCT 3. Albert Mullin asks us an edu
cational problem:
A very wise dean wishes to improve the
quality of research and education at her
university. She believes some of the
"fault" is with the department heads,
but, being wise, recognizes some of the
"fault" may be with herself. For starters,
she decides to relocate her office so as
to minimize the average distance to her
department heads. Where should she
locate her office? (For simplicity, assume
all offices are located in a plane and that
the dean cannot co-reside with any de
partment head).
OCT 4. Yogesh Gupta posted the fol
lowing problem on an electronic bulletin
board that I read:
An exploring team wants to reach a des
tination that is six days away. Each ex
plorer can carry enough provisions to
sustain one person for four days (and
the distance an explorer travels in a day
is independent of the amount of provi
sions he or she is carrying). What is the
smallest team that permits at least one
explorer to reach the destination and
permits all the explorers to return home
safely?
OCT 5. John Glenn has two cats, which
are fed a single can of cat food. The food
comes out of the can as a single intact
cylinder, and the can is then used to cut
the cylinder into two shares of food.
Where should the can be positioned to
generate equal shares?
Food
Can
Speed Department
SD 1. Phelps Meaker wants to know at
what time between two and three o'
clock the hour and minute hands are in
a symmetrical position.
SD 2. We end with a series of "chem
istry" questions from The Tech, M.I.T.'s
student newspaper:
Give the names of the following com
pounds:
I WoNoCrKrWsNoCrKrVWNoCfKrWoNeCrKrWlNoCrierWoNoCtKl
2. Be +At-
3. BotuHIJKLMnO
4.
10 TOR TOR
9. (BoNa2)l2
F.++
A30 OCTOBER 198S
Melvin Moisted, '23; May 12, 1985; Philadelphia,Penn.
Millon O. Orwin, '23; December 11, 1984; Kala-mazoo, Mich.
Lewis J. Powers, '23; 1985; Springfield, Mass.
Mrs. Alice C. Bishop, '25; Januarv 20, 1982; Cran-ford, N.J.
Kathlyn A. Carnagey, '25; October 7, 1983; Badin,N.C.
William J. Mahoney, '25; June 12, 1985; Chatham,Mass.
Vincent B. Bennett, '26; March 1985; Ipswich, Mass.
Philip A". Hendee, '26; September 25,1984; Charleston, W.V.
William H. Hoar, '26; April 27, 1985; SUver SpringMd. V BJames S. Offuti, '26; May 14, 1985; Seminole, Fla.Alfred P. Steensen, '26; March 30, 1985; Manchester, N.H.
Mrs. Edward). Poitras, '28; February 16,1985, VeroBeach, Fla.
John L. Cantwell, '29; May 8, 1985; Chicopee Falls,Mass.
Willis F. Davis, '29; April 14, 1985; Erie, Penn.Harold C. Pease, '29; May 30, 1985; St. Petersburg,
John P. Rich, '29; November 22, 1984; Nashua,N.H.
Warren W. Walker, '29; April 20, 1985; Montclair,N.J.
John F. Guinan, '30; May 18,1985; Arlington, Mass.Bertwell M. Whitten, '30; June 5, 1985; Searsport,Maine.
Alexander Cameron Crdsman, '32; 1985; Santa Ana,Calif.
Lawrence W. Grady, '32; June 6, 1985; South Yarmouth, Mass.
Samuel E. Paul, '32; May 4, 1985; Napa, Calif.
William D. Murphy, '33; May 22,1985; Annandale,Va.
John B. Smyth, '33; April 3, 1985; Peekskill, N.Y.Charles H. MacFarland III, '34; June 13, 1985;Union, Conn.
Herbert A. Morriss, Jr., '34; February 9, 1985; Manhattan Beach, Calif.
Gerald M. Reed, Jr.,'34; July 4,1985; North Quincy,Mass.
Alfred Z. Boyajian, '35; 1985; Scituate, Mass.Willard R. Crout, '35; October 1983; Pittsburgh,Penn.
Alex G. Keiller, '35; 1985; Encinitas, Calif.
Alwin R. Knoeppel, '35; April 1984; New York,N.Y.
Robert T. Sutherland, Jr., '35; May 31, 1985; SandyHook, Conn.
Lincoln P. Vennard, '35; November 1984; Burbank,Calif.
Richard M. Whitmore, '35; March 3,1985; Whiting,N.J. 6Dana Devereux, '36; May 15, 1985; New Canaan,Conn.
Robert L. Alder, '37; April 6, 1985; Garland, Tex.
Melville E. Hitchcock, '37; March 9,1985; Ivoryton,Conn.
Leslie A. Johnson, '37; April 20,1985; Sanford, N.C.
Wayne M. Pierce, Jr., '37; May 7, 1985; Orange,Conn.
George Albert Randall, '37; June 17, 1985; New-buryport, Mass.
John A. Dodge, '39; May 1985; Wayne, Penn.
William G. Hamlin, '42; June 20, 1985; Glenshaw,Penn.
Everett L. Meley, Jr., '42; April 25, 1985; Houston,Tex.
Aurelio C. Hevia, '43; May 29, 1985; Miami, Fla.Martin H. Winter, '43; February 10,1985; Paramus,N.J.
Stanley Berinsky, '44; March 4, 1985; Sunnyvale,Calif.
Frederick B. Meier, '44; 1985; Monroe, Mich.
Clarence S. Howell, Jr., '45; June 7, 1985, MercerIsland, Wash.
Donal Botway, '49; May 8, 1985.
Louis Robinson, '50; March 28, 1985; Scarsdale,N.Y.
Roland F. Beers, Jr., '51; March 20, 1985; Dorset,Vt.
Robert W. Elliott, '52; May 13, 1985; LynnfieicMass.
Morton S. Hoppenfeld, '52; March 26, 1985.
I,
Shanker R. Bagade, '55; January 1979; Cincinnati,Ohio.
Choji Nozaki, '55; April 16, 1985; Nagoya, Japan.
Leon A. Yacoubian, '55; April 1985; Damascus,Syria.
Dalton L. Baugh, Sr., '56; January 3, 1985; Malta-pan, Mass.
Charles Diebold III, '58; June 1985; Buffalo, N.Y.
Myron P. Lepie, '58; May 9, 1985; Chestnut Hill,Mass.
Raymond A. Bruce, '61; April 13, 1985; Middle-
town, N.J.
Charles C. Conley, '61; November 20, 1985; SaintPaul, Minn.
Sudhanshu K. Dikshit, '61; October 1982.
Clarke E. Swannack, '61; December 12,1984; Beloit,Wise.
Albert L. Baker, Jr., '65; March 1985; Dunedin, Fla.
Vittorio Baldini, '75; April 23, 1985; Milano, Italy.
John B. Lamb, '77; May 31, 1985; Wellesley, Mass.
David E. Miller, 79; July 1, 1985; Edwards AFB,Calif.
David J. Shapiro, '87; June 19,1985; Concord, Mass.
StemhiecherGup.
Contract research and
development In
radio frequency,
microwave and
mllimeter wave
engineering and
related areas.
RF and Microwave
Systems Design
Industrial Applications
of Microwave Power
Precision
Instrumentation
Analog and Digital
Electronics
Manufacturing
facilities available
165 New Boston Stroot
Woburn, MA 01801
Telex 948-600
(617)935-3460
Syska&HennessyInc
Engineers 5501 Qreon Valloy
Circle
Mechanical/Electrical/ Culver City
Sanitary Los Angeles, CA
90230
John F. Henrtessy '51
11 West 42nd St.
New York. N.Y.
10036
840 Memorial Or.
Cambridge, MA
02139
657 Mission St.
San Francisco, CA
94105
TedmolpgyMarketingGroup Inc
Marketing Consultantstor Technology
Companies
Extending Your
Resources
to Achieve:
» Sales Growth
• New Markets
• New Applications
• Product
Enhancements
• Profitable Pricing
Leslie C. Hruby, SM 73
F. Michael Hruby
326 Burroughs Rd.
Boxborough, MA 01719
(617)263-0643
ThomasK. Dyer,
Inc
DIVISION OF HNTB
Consulting Engineers
Rail Transportation
Thomas K. Dyer '43
1762 Massachusetts Ave.
Lexington, MA 02173
(617)862-2075
Washington, D.C.
(202) 466-7755
Chicago, IL
(312) 663-1575
Philadelphia, PA
(215) 569-1795
Venttnesand
Adventures
Start-Ups Giorgio Plcotgll,Management '67, Ph.O.
Consultation
Special Pro|ects 100 Dorado Terrace
Sen Francisco, CABridging 94112
The Technical (415) 333-S0B4and
The Organizational
The Bureaucraticand
The Organizational
Specializing In the
solution of complete.
difficult, now,
Infrequent, or unique
problems
TECHNOLOGY REVIEW A2»
Solutions
M/J 1. How does South make five clubs after an
opening lead of the »Q, and what lead sets this
contract?
♦ A 84
¥ AK 10
♦ A7654
♦ KJ
A K
V QJ9432
♦ Q9832
+ Q
* QJ10 3 2
V 865
♦ JA 107 54
♦ 9765
V 7
♦ K 10
♦ A98632
Benjamin Feinswog suggests winning the opening
lead with the ♦ K and then cashing the AK and
*J, discarding a spade. Now the V10 is ruffed and
the *A cashed. At this point the ♦ 10 is led and,
whether East ruffs or not, the ♦A provides an entry
to the ♦ A permitting a second spade discard. ThusSouth loses only one spade and one trump. Mr.
Feinswog notes that a lead of the ♦ 2 (among others)
will set the contract.
Also solved by Richard Hess (who believes that
any lead will set the contract), Matthew Fountain,
and the proposer, Doug Van Tatter.
M/J 2. A large bed of flowers and greenery is laid
out in the form of a regular polygon of N sides. A
walk composed of N trapezoidal concrete slabs sur
rounds the flower bed. A circumscribing circle pass
ing through the outer corners of the walk and an
inscribing circle tangent to the inner flats of the
trapezoids have circumferences in the ratio of al
most exactly 191:165. The total area within the outer
periphery of the walk and the area of the walk itself
are in the ratio of 4:1. Find N.
The following solution is from Richard Helden-fels:
The general equations of regular polygons and a
trigonometric relationship were all that was re
quired to obtain a direct solution. The ratio of cir
cumscribing and inscribing circles provides thefollowing:
a.tan(180/N)/a1sin(180/N) = 191/165,
where the a's are the length of the outer and inner
sides and N is the number of sides. The rano ofareas provides the following:
A./Ao = (a,/a<,)J = 3/4.
These equations can be combined and simplified toyield:
cos(180/N) = (165/191)(2/V3).
The solution is: 180/N = 4.0391°. Since the ratio ofcircles was not exact, N = 45. The ratio of 191:165
is greater than that ratio for N = 45 polygon by the
factor of 1.000048. A drawing of one trapezoidalconcrete slab follows:
26646
25 526
23 076
Also solved by Charles Freeman, Dennis White,
Frank Carbin. George Parks.Harry Zaremba. John
Prussing, John Woolston, Matthew Fountain, Mi
chael Jung. Naomi Markovitz, Richard Hess, Steve
Fcldman, and the proposer, Phelps Meaker.
M/J 3. Given a number N between 500 and 1000,
rapidly construct a scries of numbers using eight of
the nine digits from 1 through 9 once, the other
digit twice, and a few zeroes so the series totals N.
Naomi Markovitz must have spent a great deal
of time in tricky parlors:
HTU
1
2
3
4
5
6
7
8
9
HTU
1
2
3
4
5
6
7
8
9
H T 11
1
2
3
4
5
6
7
8
9
HTU
1
2
3
4
5
6
7
8
9
3
Being that as many zeros as desired may be used,
I'll worry about whether the other digits should be
in the hundreds, lens, or units column and then
affix the appropriate zeros in the presentation of the
requested series. 1 noticed in the example presented
that most of the non-zero digits are in the tens column. Indeed the sum of 10 + 20 + . . . + 90 =
450, which is the same order of magnitude as any
answer which may be requested. Therefore, my ap
proach would be to start with one "copy" of each
digit in the tens column and then to check, in the
case of each sum desired, which digits have to be
moved to the hundreds column, which to the units
column, and which has to be duplicated (and the
column in which the second copy will appear). If a
digit, a, moves from the lens to the hundreds col
umn, the sum will increase by 100a - 100a = 90a.
If a digit, b, is moved from the lens column to the
units column, the change is - 10b + b, or a decrease
of 9b. Take, for example, the case which is pre
sented in the problem: N = 642.1 start off with the
arrangement at the left of the box above. This gives
a total of 450. 1 have to add 642 - 450 = 192. To
check which digit has to be moved to the hundreds
column, I divide 190 by 90 and get 2 +. I'll move 3
to the hundreds column, because if 1 add too much,
1 can subtract the extra by moving a digit to the
units column. Now I have the second display in the
box. I've added 3 x 90 = 270, instead of the re
quired 192, so I'll have to subtract 78. To determine
which digit to move to the units column, I divide
78 by 9 and get 8 +. Again, I chose to subtract "too
much" because this problem can be solved by add
ing the second copy of one of the digits. So I'll move
the 9 into the units column. Now I have the third
display in the box. I have subtracted 9 x 9 - 81
instead of the required 78. I have to add 81 - 78
= 3. I therefore add a second 3 to the units in the
fourth display. So, the answer can be presented as:
319
20
40
50
60
70
63
642
There are, however, cases in which the procedure
must be slightly modified. The basic procedure will
work in all cases since:
500 « N « 1000
we see that
50 « N - 450 « 550.
Dividing by 90 will always give an, answer in which
a single digit must be moved to the hundreds col
umn. (Even if the division comes out even, add 1
to the quotient to determine the digit to be moved,
because eventually a second copy of one digit will
have to be added.) The second stage consists of
dividing by 9 a number from 1 to 90. Several prob
lems could occur:
(1) The number that has to be moved from the tens
AlexanderKusko, Inc.
Research,
dovolopmont and
engineering sorvlcos
In the electrical
anglnoorlng field
Specialties:
Electric power
systams
Eloctrlc transportation
oqulpmont
Electric machinery
and magnetics
Soild-stato motor
drives, roctlflors,
Invorters
Feedback control
systems
Computer applications
and modeling
Evaluation,
Investigation,
patents.
Alexander Kusko '44
161 Highland Avenuo
Needhom Heights, MA
02194
(617)444-1381
Boyleyi •' • •
Gotp.
Engineers/Architects
Complete Professional
Services:
Water Resources
Pollution Control
Architects and
Landscapa
Architects
Highways and Bridges
Dams and Reservoirs
Facilities System
Environmental Sclonce
Computer Sclencos/
CADD
Agricultural Services
Management and
Administration
Transportation
Engineering
Thomas S. Maddock '51
1501 Quail Street
P.O. Box 7350
Newport Beach, CA
92658-7350
(714) 476-3400
Airfield/Airport
Facilities
Port Facilities
Health Care Consultants
Design, Construction,
Management
Subsidiaries:
Charles N. Debos &
Assoc. Inc.
Alma Nolson Manor Inc.
Par* Strathmoor
Corporation
Rocklord Convalescent
Center Inc.
Chambro Corporation
Charles N. Debes '35
5666 Sirathmoro Drive
Rocktord, IL. 61107
TECHNOLOGY REVIEW A31
Edward RMaiden
Corp.
Builders for Industry,
Institutions, hospitals,
manufacturing plants,
government and
dovelopsra of high
technology facilities
for over 35 years
Edward R. Marden '41
Douglas R. Marden '82
260 Lincoln Street
8OSUMI, MA 02134
(617) 782-3743
Georgp A.
Roman &
Associates Inc.
Architecture Planning
Interior Design
CocnfnorciQl
industrial
Residential
Site Evaluation
Land Use Planning
Master Planning
Programming
Interior Space
Planning
College*
Hospitals
Medical Buildings
Office Buildings
Apartments
Condominiums
George A. Roman,
A.LA. '65
One Gateway Center
Newton, MA 02153
(617)332-5427
Gqldbog-Zoino &
Associates Lie
Geotechnical-
Geohydrological
Consultants
The GEO Building
320 Needham Street
Newton Upper
Falls. MA 02164
(617) 369-0050
Other Offices:
Bridgeport. CT
Vernon. CT
Tampa. FL
Manchester. NH
Buffalo. NY
Providence. Rl
D. T. Goldberg, S4
W. S. Zoino. '54
J. O. Guertln, 67
M. J. Bsrvenlk. '76
D. M. Brown. 81
M. 0. Bucknam. '81
N. A. Campagna. Jr. 67
F. W. Clark. 79
K. A. Fogarty. '81
W. E. Hadge. '79
C. A. Lindberg. 78
R. T. McGIIHvray. '68
C. A. Mendez. 85
J. D. Okun. 75
K. J. 0 Reilly. '80
R. M. Simon. 72
T. von Rosenvinge IV.
'80
W. G. Williams. '65
0. W. Wood. 76
column to the units column is no longer in the tens
column, because it has already been moved to Ihe
hundreds column. Solution: If b £ 3, move both 1
and b - 1 (or any other combination of 2 numbers
totaling b) into Ihe units column. If b - 2, instead
of moving 2 into Ihe hundreds column (for a gain
of 2 x 90 = 180). move the 3 into the hundreds
column and then 1 + 9 (or any other combination
totaling 10 other than 2 + 8) into the units column:
3x 90 -1x9-9x9 = 180
The 2 will then be available to move into Ihe units
column.
If b " 1, instead of moving 1 into the hundreds
column, move 2 inlo the hundreds column and
either 3 and 7 or 4 and 6 into Ihe units column. The
1 will then be available to be moved into the units
column.
(2) It's required to move 10 or 11 into the units
column. Solution: Move in two numbers with the
appropriate sum. There will always be a pair avail
able, because the original list (1-9) consists of four
pairs totaling 10 and four pairs totaling 11. Since
only one digit has been removed from the tens col
umn,enough choices still remain. The last stage al
ways gives a number from 1 to 9 that can be added,
so this can be accomplished by placing the appro
priate single digit in the units column.
Also solved by Mathew Fountain, Richard Hess,
Steve Fcldman, and the proposer. Lester Steffens.
M/J 4. As we generate geometric figures to repre
sent y ■=> x", we have "elements" consisting of
points, lines, faces, cubes, etc. as n increases. For
the number of points in each figure, we have (for
n > 0) P = 2". Derive the number of lines and faces
in a five-dimensional hyper-cube.
Charles Sutton solved a generalization of M/J 4:
We may consider the generation of the sequence of
n-dimensional cubes, d units on a side, as follows:
A cube of any number of dimensions, moved a dis
tance d in a direction perpendicular to all its di
mensions, will generate a cube of one higher
dimension. If one starts with a zero-dimensional
cube (a point), this procedure will generate, in
succession, line segments, squares, cubes, and so
on into cubes of higher dimensions (hypercubes).
Nole that an n-dimensional cube will have as com
ponents all cubes of lower dimensions, from zero-
dimensional points up to (n - l)-dimensional
cubes. Now when an n-dimensional cube is moved
a distance d in Ihe direction of the next higher di
mension to generate an (n + l)-dimensional cube,
its r-dimensional component cubes (0 s i < n) will
necessarily generate (r + l)-dimensional compo
nents of Ihe (n + l)Klimensional cube. It is clear
that Ihe number of points will double, since there
will be just as many points in the final position of
the cube as in Ihe initial position, and hence the
number of points (zero-dimensional components)
in an n-dimensional cube will be 2". When an r-
dimcnsional component (r s> 1) of an n-dimensional
cube moves a distance of d, there will be just as
many such components in Ihe final position as in
Ihe initial position, but there will be additional r-
dimensional components generated by the motion
of (r - l)-dimensional components. Hence the
number of r-dimensional components in an (n +
l)-dimensional cube can be obtained by adding the
number of (r - l)-dimensional components to twice
Ihe number of r-dimensional components in an n-
dimensional cube. We can use this to complete the
array at the top of the next column, giving the lower
dimensional components for cubes of each number
of dimensions. Filling in l's in the diagonal (since
an n-dimensional cube can be considered to have
itself as a component) and powers of two in the left
hand column, we can continue by adding twice the
value of any entry to its left hand neighbor and
writing the result below the doubled entry. The con
struction of this array is reminiscent of Pascal's tri
angle, and in fact the numbers in the rows can be
seen to be the coefficients in the expansion of (2x
+ 1)". And the answer to Winslow Hartford's ques
tion in M/J4 is that the numbers of lines and faces
in a five-dimensional hypercube are both 80.
Mr. Sutton then adds the following remarks:
I have always been intrigued by the geometry of
Dlncnslon of components
2 3 4 5 6 7
Dloenslon
of cube
0;
2
3
4
s
6
r
l
2
4
8
16
32
64
128
1
4
12
32
80
192
448
1
6
24
90
!40
672
1
8
40
160
160
1
10
160
290
1
12
84
higher dimensional figures, and recall having read
someplace that there are six regular four-dimen
sional solids but that only the analogues of the cube
and tetrahedron exist for all higher dimensions.
That set me to wondering what the triangular array
I had obtained in my solution of M/) 4 would look
like for the equilateral triangle, regular tetrahedron
scries. Actually it's fairly easy, since to step up one
dimension you need only to find a point in the next
higher dimension that is the same distance from Ihe
set of equidistant points in the lower dimensions.
The n-dimensional analogue of Ihe tetrahedron
would have n + 1 vertices, and the number of its
r-dimensional components (each having r + 1 ver
tices) would be the number of combinations of n +
1 things taken r + 1 at a time. Binomial coefficients!
And the numbers in the horizontal row of the tri
angular array corresponding lo the n-dimension.il
analogue of the tetrahedron turn out to be the co
efficients of (x + 1)"'' - x"*1. As a check, for n =
3, a tetrahedron has four vertices, six edges, four
faces, and one tetrahedron, coefficients of (x + If*1- x"*1.
Also solved by Avi Omstein, Charles Freeman,
Dennis White, Matthew Fountain, Richard Hess,
and Winslow Hartford.
M/J 5. Take the letters in the first half of alphabet
in order. Place the A on the table. Place the B next
to the one of the four sides of the A. Place the C
next to one of the six sides at the AB (or BA) pair.
Then add the D and so on. If you do this correctly,
when you reach the letter M you will have created
a crossword-puzzle matrix of complete common
English words, no proper names, no foreign words,
and no acronyms or abbreviations. Having solved
the problem as posed, can you add one or more
letters lo the A-M set and still retain complete
words?
Both Rik Anderson and Matthew Fountain found
the identical solution—one that goes up to P. The
key to extending past M is the word "knop." It
would be hard to go further since adding Q would
require U and hence R S T.
H
BACK
FED N
JIG
L
H
0
P
Also solved by Harry Zaremba, M. Arch, Naomi
Markovitz, Richard Hess, and Steve Fcldman.
Proposers' Solutions to Speed Problems
SD1. 2:46 2/11
SD2.
1) Polywannacracker
2) Polar bear
3) Barium Coldwater
(barium gold H-lo-O)
4) Hi-O, Silver
5) A dozen bananas
6) Paramedics
(or paradox)
7) Metaphysics
8) Orthodox
9) Methyl ethyl
chicken wire
10) Transistor
11) Cis-boom-bah
12) FORTRAN 4
13) Mercedes benzene
14) Ferrous wheel
SEND PROBLEMS. SOLUTIONS.
AND COMMENTS TO ALLAN /.
COTTUEB. -67. ASSOCIATE
RESEARCH PROFESSOR AT
THE COURANT INSTITUTE OF
MATHEMATICAL SCIENCES.
NEW YORK UNIVERSITY. 2S1
MERCER ST.. NEW YORK.
N.Y.. 10012.
A32 OCTOBER 1985
PUZZLE CORNER
ALLAN J. GOTTLIEB
Where Squares
and Office Parties Mix
Jim Landau sent me quite an unusual
package. I have often received mail
marked fragile, but never before one
mailed indestructible. Inside was a rope
tied into a special knot and an accom
panying problem. Although I do not as
yet see how to present the problem
without having each issue of Technology
Review include an attached rope, I have
had fun with the problem myself. I
should also remark that this indestruc
tible parcel seemed to get considerably
better treatment en route than many un
marked packages I have received. Per
haps stores should try sending crystal
in packages marked indestructible.
On a personal note, I am pleased to
report that I have received tenure at
NYU.
Problems
N/D 1. We begin with a computer-re
lated problem from Al Weiss, who is
given the coordinates of four points and
wants to determine if they form the four
corners of a square. No assumption can
be made about the order in which the
points are presented. For example, a
square might occur as:
(8,8), (3,5), (2,4), (4,9).
Mr. Weiss seeks not just a solution, but
an elegant algorithm.
N/D 2. Oren Cheyette offers us a sea
sonal problem. For an office party, each
person is supposed to bring a gift for
someone else. The recipients are as
signed to givers by writing each person's
name on a slip of paper, putting the slips
in a hat, and having everyone draw a
slip. Obviously, it's no fun if someone
draws his own name. What is the prob
ability that in an office of n people, no
one draws his own name?
SEND PROBLEMS, SOLU
TIONS, AND COMMENTS TO
ALLAN I. GOTTLIEB. '67. THE
COURANT INSTITUTE, NEW
YORK UNIVERSITY, 251 MER
CER ST., NEW YORK, N.Y.
10012.
(notN/D 3. Given a square matrix A
necessarily invertible), satisfying
AA = AA',
where the prime signifies transpose op
erator, Howard Stern wants you to show
that
A = A'
using matrix operation only, i.e. without
using normed algebras and approxi
mating A by an invertible matrix.
N/D 4. Walter S. Cluett asks: What is the
lowest number of current U.S. coins (1
cent through $1.00) for which there is
no combination of coins that will equal
in value a single coin? How many such
quantities are there under 100?
N/D 5. Ronald Raines wants you to find
functions f and g satisfying
f[f(x)] = x
glgMl = - xfor all real values of x.
Speed Department
SD 1. Phelps Meaker has a pan with
perpendicular ends and sloping sides. It
is two inches deep and measures 8" x
10" on the bottom. What is the slant
height of the sides if the capacity is 200
cu. in.?
SD 2. Jim Landau wants to buy a solid-
state digital clock with a 12-hour LED
display and wonders at what time
would the largest number of LEDs be
on? The smallest number?
Solutions
JUL1. Find the smallest prime number that contains
all 10 digits.
Dan Schmoker combined analysis with some
computer work to obtain the following solution:
The smallest 10 digit number with all digits different
is 1,023,456,789. The sum of the digits in this num
ber is 45 which is divisible by 3; hence the number
is divisible by 3 and is not prime. In addition, any
rearrangement of this 10-digit number would also
be divisible by 3 and not prime. It follows therefore,
that the smallest prime number that contains all
digits must be an 11-digit number. The extra digit
cannot be 0, 3, 6, or 9; otherwise the new number
and any number formed by rearranging the digits
would be divisible by 3. and hence not prime. The
smallest number which contains all digits and could
RobertF. RoweConstruction
Consultant
Robert F. Rowe '48
Suite I-S32
Control Data Business
and Technology
Center
Your building
program should
be enjoyable.
II it's not.
lei us smooth
out Iho process.
35 years'
experience in
industrial,
commercial, and
institutional
construction.
701 East Bay St.
Charleston, S.C. 29403
(803) 722-8610
Syska&HermessyInc.
Engineer!
Moctunlcal/EIoctrical'
Sanitary
John F. Hennossy '51
11 West 42nd St.
New York. N.Y.
10036
840 Memorial Dr.
Cambridge. MA
02139
657 Mission St.
San Francisco, CA
94105
5901 Green Valley
Circle
Culver City
Los Angeles, CA
SO230
System
ArchitectsInc.
Microcomputer communications
consultants
Authorized representatives:
IBM Personal Computer and XT
DEC Professional Compulor
HP 9000 Scientific Computer
Consultants lor:
Local area networks
Micro to mainframe
communications
Scientific and business
systems development
Constantino I. Photopoutos. '77
212 East 47th Street
New York. N.Y. 10017
(212)644-6972
TECHNOLOGY REVIEW A23
possibly be prime is therefore the 11-digit number10,123,456,789. The prime lest should be started
with this number. The numbers to be tested forprime are generated by rearranging the 11-digit
number, minimizing the change in value by switch
ing the right-hand digits first. The prime test is done
by first generating a list of prime numbers. On my
machine with 48K of memory the biggest list of
prime numbers I could generate had a total of 9000
numbers with the highest prime number being
93001. By dividing with this list of prime numbers,the number 10,123,457,689 (the seventh number
considered) was identified as a possible prime num
ber. At this stage 1 began dividing by all odd numbers greater than 93001, continuing until the divisor
exceeded the square root of 10,123.457,689. This last
test only took a few minutes and showed that
10,123,457,689 was indeed prime.
Also solved by Donald Savage, Henry Hirsch-land, Jonathan S.idick, N.C. Strauss, Ned Staples,P.V. Heftier, Richard Hess, and the proposer, Matthew Fountain.
JUL 2. The original Tower of Hanoi, first described
in 1883, consisted of 64 golden disks, each of a dif
ferent diameter, stacked according to size, with the
TechnologyMmketingGroup Inc
Marketing Consultants
for Engineered
Products
Our CUortts Use Us To:
• Plan New Products
•nd Marksti
• Reduce Soiling Cost*
e Find Growth Opportunities
Leslie C. Kruby, SM 73. Sloan School
F. Michael Hruby
77 Great Rood
Suits 202
Acton, MA 01720
(817)2634648
ThomasK-Dyer,
Inc
DIVISION OF HNTB
Consulting Engineers
Hall Transportation
Thomas K. Dyer '43
1762 Massachusetts Ave.
Lexington, MA 02173
(617) 862-2075
Washington, D.C.
(2021 466-7755
Chicago, IL
(312) 6$3-157S
Philadelphia, PA
(215) S69-1795
smallest on top. The tower is to be restacked on one
of two additional sites, moving one disk al a time
off the top of one stack either to an empty site orto the top of the slack on one of the other sites,
without ever placing a larger disk on a smaller. The
original problem was to find the minimum number
of moves to transfer the enlire stack. The new prob
lem is to calculate the location of each disk after
1,001 moves have been made using the optimum
transfer procedure.
Rik Anderson found it easier to take a gianl step
forward and then 23 small steps back:
Identify the original stack as position A. If disk 1 is
initially moved to position C, Ihen disk 2 will ini
tially be moved to position B al move 2, disk 3 will
go to C al step 4, disk 4 to B al step 8, etc. Disk n
will first move at move 2'" ", to position B if n is
even, to C if n is odd. Following this rule, the first
time disk 11 moves is at move 2'° = 1024, lo position
C. At the previous move, number 1023, disks 1-10
would have reached position B. In getting lo this
position, disk 1 is at any position for two moves,
disk 2 moves every 4th move, disk 3 ever)18th move,
and disk 4 every 16th. Disk l's moves are in reverse
sequence (A,C,B,A,C,B), as are all odd-numbered
disks, while even-numbered disks visit the posi
tions in sequence (A,B,C,A,B,C). Working these
patterns backwards from step 1023, we would find
at slep 1001:
Position A - disks 1, 4, and 11 to 64
Position B • disks 2, 3, and 6 to 10
Position C - disk 5.
Also solved by Dennis While, John Prussing,
Matthew Fountain, Ned Staples, Richard Hess,
Winslow Hartford, and the proposer, Lester Stef-
fens.
JUL 3. Define an n-triangle to be a collection of n(n
+ l)/2 points regularly spaced into the shape of an
equilateral triangle with n points on a side. Define
a 3-line to be a line segment connecting exactly three
adjacent points parallel to a side of an n-triangle.
(The three adjacent points are said to be "covered"
by the 3-line). For what values of n can all points
of an n-triangle be covered by non-intersecting 3-
lines?
This question appears to be difficult, and I con
sider the problem to be still open Harry Zaremba
shows that if one drops Ihe requirement that 3-lines
are parallel to a side of the triangle, then a solution
exists for all
n = 9m and n = 9m - 1.
Dennis White, Richard Hess, and Winslow Hartford
express the belief that (with the parallel require
ment) no solutions are possible. Mr. Hartford notes
thai to have a multiple of three points, we need
n = 3m or n = 3m - 1.
and indicates that an inductive proof should be pos
sible.
JUL 4. A manufacturer makes all possible sizes of
brick-shaped blocks such that the lengths of the
edges arc integral multiples of the unit of length,
and that the number of units in the total length of
the twelve edges of the block is equal lo two-thirds
of the number of units of volume in the block What
sizes does he make?
The following solution is from Howard Stern:
AB:CD x E = FG:H1
Since the smallest value that AB can assume is 12,
it follows that E cannot be larger than 4. It also
cannot be 1 because then the times would be the
same. Therefore, E must be 2, 3 or 4. This forces A
to be 1. In addition, C, F and H must be 2, 3, 4 or
5 for the times to make sense. Since A = 1, E cannot
be 2 because then F would also have lo be 2. E
cannot be 4 either because then B would have to be
2 or 3, and it would be impossible for C, F and H
to be 2, 3 or 5. Therefore, E must be 3. This forces
C, F and H to be 2, 4 or 5, and D and I must be 6,
7, 8 or 9. From Ihe multiplication by 3, the only
possibilities are D = 6 and I = 8, or D = 9, and I
- 7. These restrictions leave only a few feasible
times to try and the only one that works is: 18:49
x 3 = 56:27.
Also solved by Avi Omstein, Dennis White,
Frank Carbin, Harry Zaremba, Henry Hirschland,
Matthew Fountain, P. Michael Jung, P.V. Heftier,
Richard Hess, Steve Feldman, Winslow Hartford,
and the proposer, Phelps Meaker.
JUL 5. An ideal pulley system supports a bucket of
water on one rope and a monkey on the other. The
bucket and monkey are in static equilibrium and at
the same vertical level. Suddenly the monkey began
climbing up its rope. Describe the motion, if any,
of the bucket.
Everyone agrees that the bucket follows the mon
key exactly, independent of the mechanical advan
tage of the pulley system. The argument goes as
follows: The monkey's change of state from rest to
an upwards velocity implies acceleration which re
quires force. The reaction force is tension xT in Ihe
monkey's rope (where x is the pulley's mechanical
advantage with respect to the monkey). The dy
namical response of the bucket to the force xT is
identical to the monkey's, so that the bucket and
the monkey accelerate upwards together. After the
monkey's acceleration stops, it continues to climb
al some velocity v, and so does the bucket.
Solutions were received from Richard Hess (who
attributes the problem to Lewis Carroll), William
Moody (who first heard the problem 58 years ago
in his 8.01 freshman physics course, taught by Prof.
"Hard-Boiled" Lewis Young), Ronald Martin,
Winslow Hartford, Harry Zaremba, Dennis White,
and the proposer, Bruce Calder.
Better Late Than Never
1980 FEB 3. Warren Himmelberger suggests the
solution (33231302212011003)
1983 APR 4. Dick Allphin reports that a similar
problem has appeared in The New York Times. Fur
thermore, Mr. Allphin believes that, when travell
ing in the rain, "for minimum soaking it pays to
make haste."
Y1984 Rik Anderson has noticed that 16 = 41OT.
1984 APR 2. John Colcman has responded.
1985 F/M 3. John Smith found that x = (sin 60)/(sin
75)a.
F/M 4: Phelps Meaker noticed that this problem was
also F/M 4 in 1984. An unintentional yearly problem!
APR 1. Michael Jung has responded.
APR 4. Michael Jung has responded.
APR 5. Frank Carbin has responded.
M/J 4. Jim Landau has responded.
July SD 2. Ben Fcinswog wants to play the VQ
before the 4K and gives a try for the contract, if
West is void in hearts.
Proposers' Solutions to Speed Problems
SD1. 2(2)5.
SD 2. 10:08 (or 10:08:08 if the clock has seconds).
1:11 (or 1:11:11).
A24 NOVEMBER/DECEMBER 1985
Edward
MaidenGup.
Buildars lor Industry,
Institutions, hospitals,
manufscturlng plants,
government and
developors ol high
technology (acuities
lor over 35 years
Edward R. Maiden '41
Douglas R. Harden '82
280 Lincoln Street
Boston, MA 02134
(617) 782-3743
Nelson,Goulsan &Associates,
Inc
Professional Staffing
Consultants
Contract Engineering
Services
Professional
Recruiting Services
Technical Personnel In
All Disciplines
333 W. Hampden Ave.
Suite 507
Engtewood, CO 80110
(303) 761-7680
Other offices In
Albuquerque,
Coloffldo Springs,
Oallas & Seattle
Paul R. Coulson
PE, '43
President
PiccagH&Associates
Corporate
Health Care
Consultants
Giorgio Piccagll,
Ph.D.. '67
100 Dorado Terrace
San Francisco, CA
94112
(415)333-5084
PUZZLE CORNER
ALLAN J. GOTTLIEB
Great Climbing Monkeys!
This past week I was quite ill with a flu-
like disease that caused me to miss my
first three days of work since arriving at
NYU more than five years ago. And
guess who strolled into the lab to inter
view our research group while I was out:
the CBS Evening News! I guess it just
doesn't pay to get sick so often.
I am sorry to say that, for some unex
plained reason, two of the October so
lutions given in the February/March
issue were unattributed. Somehow be
tween my original manuscript and the
final column, credits to Harry Zaremba
and David Griesedieck for OCT«2 and
OCT 4, respectively, were omitted. I
apologize for the error.
Finally, I would like to acknowledge
a touching letter from one of the most
active contributors to Puzzle Corner
who explained, in a warm and personal
way, why his activity would have to de
crease. This column is dedicated to JohnRule.
Problems
JUL1. For our computer problem of the
month, Matthew Fountain wants you to
find the smallest prime number that
contains at) 10 digits.
JUL 2. Lester Steffens asks us to answer
a new question about the widely known
Tower of Hanoi problem:
The original tower, first described in
1883, consisted of 64 golden discs, each
of a different diameter, stacked accord
ing to size, with the smallest on top. The
tower is to be restacked on one of two
additional sites, moving one disc at a
time off the top of one stack either to an
empty site or to the top of the stack on
one of the other sites, without ever plac
ing a larger disc on a smaller. The orig
inal problem was to find the minimum
number of moves to transfer the entire
stack. The new problem is to calculate
the location of each disc after 1,001
moves have been made using the opti
mum transfer procedure.
JUL 3. Charles Bostick has some points
that need to be covered:
Define an n-triangle to be a collection
of n(n + l)/2 points regularly spaced
into the shape of an equilateral triangle
with n points on a side. Define a 3-line
to be a line segment connecting exactly
three adjacent points parallel to a side
of an n-triangle. (The three adjacent
points are said to be "covered" by the
3-line). For what values of n can all
points of an n-triangle be covered by
non-intersecting 3-lines?
JUL 4. Here is one from a batch John
Rule sent me in 1974 and I have been
periodically milking ever since:
A manufacturer makes all possible
sizes of brick-shaped blocks such that
the lengths of the edges are integral mul
tiples of the unit of length, and that the
number of units in the total length of the
twelve edges of the block is equal to two-
thirds of the number of units of volume
in the block. What sizes does he make?
JUL 5. Here is some monkey business
from Bruce Calder:
An ideal pulley system supports a
bucket of water on one rope and a mon
key on the other. The bucket and mon
key are in static equilibrium and at the
same vertical level. Suddenly the mon
key began climbing up its rope. Describe
the motion, if any, of the bucket.
SEND PROBLEMS. SOLUTIONS.
AND COMMENTS TO ALLAN I
GOTTLIEB. '67. ASSOCIATE
RESEARCH PROFESSOR AT
THE COURANT INSTITUTE OF
MATHEMATICAL SCIENCES,
NEW YORK UNIVERSITY. 251
MERCER ST.. NEW YORK.
N.Y.. 10012.
A22 JULY 1985
Ralph E. Brown, '26; December 24. 2984; Fairless
Hills, Penn.
Robert H. Clarke, "26; November 4. 1985; Dayton.
Ohio.
Thomas J. Eaton, '26; April 17, 1984; Tucson, Ariz.
William M. Smith, '26; 1985; West Hollywood. Fla.
Mrs. William Wraith, Jr., 76; 1983; Tucson. Ariz.
William T. Corey, '27; December 30, 1983; Garden
City, N.Y.
Frederick J. Hooven, '27; February 5, 1985; Nor
wich. Vt.
Roger M. Pierce, Sr., '27; January 10, 1985; West
Brookfield, Mass.
Charles M. Anderson, '28; December 1984; Closter,
Miss.
Albert E. Beitzell, '28; February- 3. 1985; Bangor,
Maine.
Harold A. Harrington, '28; July 16. 1984; Cam
bridge, Mass.
Albert H. Shedd, '28; 1985; Molalla. Ore.
Richard S. Smith, '28; September 24. 1984; Sacra
mento, Calif.
Marshall H. Fay, '29; February 8, 1985; Port Wash
ington, N.Y.
Harold H. Theiss, '29; 1983.
Edward D. Thomas, '29; February 1, 1985; Foxboro,
Mass.
Claude C. Cash, '30; July 25, 1984; Hohokus, N.J.
Richard B. Ellis, '30; March 5, 1985; Athol, Mass.
John C. Larkin, '30; February 3, 1985.
Fred L. Markham, '30; 1985; Provo, Utah.
William Wallace McDowell, '30; March 2.1985; Na
ples. Fla.
Myron T. Smith, '30; February 18. 1985; South
Casco, Maine.
William, E. Yelland, '30; January- 27. 1985; South-
borough, Mass.
Paul A. Davis, '31; December 1, 1983; Vero Beach,
Fla.
Herman H. Ferre, '31; February 1985; Ponce, PR.
Robert M. Kelly, '31; January 30, 1985; Orleans,
Mass.
Manuel Schivek, '31; July 31. 1984; Randolph,
Mass.
Charles R. Wood, '31; September 9. 1984; Kennett
Sq., Penn.
Earl F. Anderton, '32; November 27,1984; Bellevue.
Wash.
Mrs. Amy V. Higgins, Jr., '32; 1982.
G. Edward Nealand, '32; March 19,1985; Sandwich.
Mass.
George W. Palmer, '32; December 20, 1984; Fal-
mouth Foreside, Maine.
Sterling N. Slockbower, '32; 1985; South Plainfield,
N.J.
Walter R. Duncan, '33; January 20,1985; Rosemont,
Penn.
Henry E. Kiley, '33; September 1984; Chatham, N.J.
Harold H. Okasaki, '33; 1984; Gardena, Calif.
John G. Trump, '33; February 21, 1985; Winchester,
Mass.
Arthur E. Byerlein, '34; 1982; Tucson, Ariz.
Robert M. Elliott, '34; December 14, 1984; Norwich,
Conn.
David L. Foulkes, '34; January 23, 1985; Oakland,
Calif.
Alton C. Garland, '34; January 1, 1985; Sandwich.
Mass.
Horace A. Giddings, '34; December II. 1984; Day-
tona Beach, Fla.
James R. Higgins, '34; October 18. 1984; Kansas
City, Mo.
William W. Buechner, '35; March 12. 1985; Arling
ton, Mass.
George R. Bull, Jr., '35; December 9,1984; Wayne,
Penn.
Luis A. Dastas, '35; 1985; Miami, Fla
Richard K. Anderson, '36; 1985; Sumter, S.C.
Charles L. Austin, '36; 1984; Sun City, Ariz.
William Fingerle, Jr., '36; December 6, 1984; Old
Greenwich, Conn.
Carlyle W. Jacob, '36; February 18, 1985; Quincy,
Mass.
George B. Payne,. '36; June 28, 1984; Badiff. Tex.
G. Elliott Robinson, '36; March 30. 1985; Hanover.
Mass.
Annis G. Asaff, '37; March 1983; Lincoln, Mass
Robert L. Carlisle, '37; 1985; Bayside, N.Y.
Peter Kolupaev, '37; January 24,1983; Philadelphia.
Abraham Schwartz, '37; February 4. 1984; Engle-
wood, N.J.
Arthur B. Savel, '38; January 10, 1985; Brookline,
Mass.
Peter E. Kyle, '39; December 6, 1984; Northfield,
Vt.
Parks R. Toolin, '39; 1985; Pittsburgh, Penn.
William C. Walker, '39; January 27, 1985; Lexing
ton, Mass.
Bernard Carver, '40; 1985; Winlhrop, Mass.
Nils M. Rosenberg, '40; January 20, 1985; Seattle,
Wash.
Louis Strymish, '40; 1985; Newton Center, Mass.
Nicholas Williamson, '40; January 20, 1985; Pocas-
set. Mass.
Preston R. Glading, '41; December 16, 1984; Bar-
rington, R.I.
Peter Homack, '41; December 1984; Palm Beach, Fla.
Samuel L. Solar, '41; June 1984; San Jose, Calif.
Robert W. Curtis, '42; November 27, 1984; Arling
ton, Va.
David F. Kinert, '42; 1985; Burlingame, Calif.
John W. McNall, '42; July 10, 1984; Greensburg,
Penn.
Maynard D. Lee, '44; January 3, 1985; Kittery,
Maine.
Leonard T. Loforese, '44; December 13, 1984;
Greenwich, Conn.
Hugh M. Jansen, Jr., '45; February 25, 1984; At
lanta, Ga.
Edward V. Oxenford, '45; February 4, 1985; Buenos
Aires, Argentina.
Edward J. Fradkin, '46; February 10. 1985; New
York, N.Y.
Frederick M. MacDonald, '46; January 8, 1985;
Hyannis, Mass.
Walter H. Amadon, '48; October 26,1984; Sudbury,
Mass.
James Dugundji, '48; January 8,1985; Los Angeles,
Calif.
Peter W. Johnson, '48; 1985; San Jose, Calif.
Mrs. Paul Dulaney, '49; January 1985; Glade Spring,
Va.
William Haddon, Jr., '49; March 4, 1985; Bethesda,
Md.
Worley B. Lynn, '49; 1984; Vian, Okla.
Eugene A. Morgan, '49; March 14, 1985; Bedford,
N.H.
Axel Erik Nygrcn, '51; January 8, 1985; Fagersta,
Sweden.
Sergej Zezulin, '52; February 2.1985; Sea Cliff, N.Y.
William E. Sollecito, '53; November 2, 1984; Syra
cuse, N.Y.
John B. Padgett, Jr., '54; July 16.1984; Palos Verdes,
Calif.
David R. Wones, '54; December 1984; Blacksburg,
Va.
William B. Banks, 55; February 23, 1985; Port Or
ange, Fla.
Thomas C. Wood, '55; April 19, 1984; Manhattan
Beach, Calif.
Robert R. Pollard, '56; January 29, 1985; Hollis,
N.H.
Harry M. Salesky, '57; 1985; Tiburon, Calif.
Robert Van Benschoten, '57; March 15, 1984; Liv
ingston. N.J.
Donald Daryl Wyckoff, '58; January 20, 1985; Mar-
blehead. Mass.
Bruce R. Hayworth, '59; January 20, 1985; Poway,
Calif.
Maciej James Achmatowicz, '64; October 28. 1984;
Ashburn, Canada.
Ralph W. McKenney, Jr., '66; January 18, 1985.
Stanley D.Derbin, '69; November 29, 1984; Dan-
bury, Conn.
Kenneth R. Britling, 71; February 18, 1985; Dover,
Mass.
Omer S. Kaymakcalan, '75; February 22, 1985; De-
wilt, N.Y.
David A. Anderson, 76; February 22, 1985; Ha-
worth, N.J.
Wayne E. Matson, '77; February 1985; Boston.
Randall G. Chipperfield, '85; Cambridge, Mass.
Karl N. Horita, '85; January 25. 1985, Brookline.
Mass.
AlexanderKusko, lite
Research, Alexander Kusko '44
development and
onglnoorlng sarvlcos 161 Highland Avenue
In the electrical Noodham Heights, MA
engineering flsld 02194
(617) 444-1361
Spoclaltles:
Electric power
systems
Electric transportation
equipment
Electric machinery
snd msgnotlcs
Solid-state motor
drives, roctlNors,
Invertors
Feedback control
systems
Computer applications
and modeling
Evaluation,
Investigation,
patents.
LEA Group
UNENTHALEISENBERQ
ANDERSON, INC.ENGINEERS
Building Design
Environmental
Engineering
Slto/Clvll Design
Roofing Technology
Consultants to
Industry, commorce,
government and
Institutions.
75 Kneoland Stroot
Boston, MA 02111
(617) 4264300
New York, NY
(212)509-1922
Eugene R. Elsenoerg
«
Louis Rexroat
Anderson 'SO
William S. Hartley '52
David A. Peters 77
MSCE
Vlesturs H. Ule '78
LaidElectricCompany
Inc.
Electrical contractors
to the nation slnco
1835
Headquarters:
45 Rockelellor Plaza
New York, N.Y. 10111
Offices In 16 principal
cities throughout the
U.S. and Puerto Rico
Boston Office:
86 Coolldge Avo.
Watertown, MA 02172
(617) 926-5500
Aloiander W. Mortal. Jr.
TECHNOLOGY REVIEW A21
Speed Department
SD 1. Phelps Meaker is building astraight sidewalk 54'8" long, using 40
cast-concrete slabs each formed as an
equilateral triangle. To square off the
ends, he has two extra 30-60-90 half-slabs. What is the width of the walk?
SD 2. A bridge quickie from Doug VanPatter:
Your
A
¥
♦
*
A
A
K
A
hand:
K10
QJ1072
Jio
Dummy:
AJ53
V 965
♦ AJ74
* 062
Your contract is six hearts. West leads alow club. You put in the #Q, which is
covered by the *K and *A. Can you
find a line of play that just about guarantees success?
Solutions
FEB/MAH 1. Management meetings are scheduled
on the second Thursday of each month, adminis
trative conferences are the third Friday, and work
units have a seminar on the first Monday. Derive
an alogrithm which will generate a date given the
year, month, day-of-week, and ordinal week within
the month. For example, if a meeting were scheduled for the third Friday of August 1984, the algo
rithm would return "August 17." Note that a
meeting on the fifth Tuesday in March would be
fine for 1983, 1985, and 1986 but not for 1984 (thereare only four Tuesdays in March of 1984). In this
case the algorithm could return January 0.
The following solution is from Al Weiss, who also
notes that there are only four Tuesdays in March
in 1985 and 1986:
Programmers, like all good craftsmen, have their
own bags of tools. One tool that has been in my
bag for a long time is a routine that will convert a
date into the number of days since November 24,
-4713 (this is not the same as 4713 B.C., since there
was no year zero). The number generated by this
routine is called the Julian day. It is similar to a
"shop date," which is the number of days since the
beginning of the year. This number itself is not too
useful, but it can be used to calculate the number
of days between two dates. There is also a com
panion routine which converts the number back
into a date. These two routines can be used of find
the date 90 days from today. Using these two rou
tines it is possible to solve Alfred Anderson's problem. My procedure is as follows:
1. First we determine the Julian Day of the 1st of
the month desired (assuming we are looking for the
third Friday in July 1985, this would return the number 2446248).
2. We then determine what day of the week this is
(in this case the program returns a 1 telling us that
this is a Monday). Next we find the date of the firstFriday (the computer tells us that this is the 5th).
3. Then we calculate the date of the 3rd Friday (inthis case the 19th).
4. Finally the program validates the answer by cal
culating the Julian day of this dale and then reconverting the Julian day back to a real date. If themonth of this new date is the same as the original
month, we have a solution. As an example of an
impossible date: if we were looking for the 5th Fri
day in July, the computer would calculate.it to bethe 33nd of July. When this is converted to a Julian
day and then back to a real date it comes back as
August 2nd. The computer recognizes that August
is not the month requested and sets the date to zero.Mr. Weiss submitted the program described
above, but inadequate space prevents us from reprinting the program and we apologize. A copy can
be had by return mail from the Review.
Also solved by Frank Carbin, Harry Zaremba, JimLandau, John Patterson, Matthew Fountain, Robert
Slater, Winslow Hartford, and James Abbott (whoalso included a Tl-59 program card containing hissolution.)
FEB/MAR 2. Two coins, loosely coupled, are flipped
simultaneously such that if either one is heads, theother has probability 7/8 of also being heads, but ifeither one is tails, the other is equally likely to be
either heads or tails. Find the probability of eachindividual coin turning up heads, and the probability of their both being heads simultaneously (or
prove that the problem statement and data are inconsistent).
Michael Tamada found a couple of solution techniques for the loosely coupled coins:
Each coin has probability of .8 of being heads. Theprobability that both are heads is .7:
Coin I
I have two ways of deriving the answer, one usinga "contingency table" approach and one using a"conditional probability" approach.
Contingency Table Approach:
We wish to find the unknown probabilities a, b, c,and d:
H
T
Co
H
.7
.1
.a
in 2T
,|
.1
.2
.8
.2
Coin 2
Coin I "-
H
o
c
T
b
d
We know that a + b + c + d = l. If Coin 1 is tails(i.e., if we're in the second row, which contains c
and d), we are told that Coin 2 has equal probabilities of being heads or tails. In other words, c/(c +
d) = 1/2. Similarly, if Coin 2 is tails, we are told
that b/(b + d) = 1/2. These two equations tell usthat b = c = d, so we know that a + 3b = 1. We.are told that if Coin 1 is heads, then Coin 2 has a
probability of 7/8 of being heads. In other words,
a/(a + b) = 7/8. Substituting a = 1 - 3b in theabove equation, we get (1 - 3b)/(l - 3b + b) =
7/8, or b = .1. Since b = c = d, we know c = d
= .1 and thus a =.7.
Conditional Probability Approach:
Let "HI" and "Tl" stand for the probability that
Coin 1 is heads or tails respectively. Obviously HI
= 1 - Tl (and H2 = 1 - T2 for Coin 2). Letp(Hl:H2) stand for the probability of event HI given
that H2 occurs. We are told that p(Hl:H2) = 7/8
and p(H2:Hl) = 7/8. Similarly, we are told that
p(Hl:T2) = 1/2 and p(H2:Tl) = 1/2. The uncondi
tional probability of any event is equal to the sum
of its conditional probabilities (weighted by the
Geostge A.
Roman &Associates Inc
Architecture Planning
Interior Design
Institutional
Commercial
Industrial
Residential
Site Evaluation
Land Use Planning
Master Planning
Programming
Interior Space
Planning
Colloges
Hospitals
Medical Buildings
Office Buildings
Apartments
Condominiums
George A. Roman,A.I.A. '65
One Gateway Center
Newton, MA 02158
(617) 332-5427
RobertRRoweCbnshtidianConsultant
Robert F. Rowe '48
Suite 1-532
Control Data Business
and Technology
Center
Your building
program should
be enjoyable.
II it's not.
lot us smooth
out the process.
35 years'
experience In
Industrial,
commercial, and
institutional
construction.
701 East Bay St.
Charleston. S.C. 29403
(803) 722-B610
SteinhtecherCrap.
Contract research and
development In
radio frequency,
microwave and
milimeter wave
engineering and
related areas.
RF and Microwave
Systems Design
Industrial Applications
of Microwave Power
Precision
Instrumentation
Analog and Digital
Electronics
Manufacturing
facilities available
185 New Boston Street
Woburn, MA 01801
Telex 948-600
(617)935-8460
TECHNOLOGY REVIEW A23
Syska&HermessyJnc
Engineers
Mechanlca('Electrical/
Sanitary
John F. Honnessy '51
11 West 42nd St.
New York, N.Y.
10036
840 Memorial Or.
Cambridge, MA
02139
657 Mission St.
San Francisco, CA
94105
SS01 Green Valley
Circle
Culver City
Loa Angelas, CA
90230
TechnologyMadceting
Group Inc
Marketing Consultants
for Technology
Companies
Extending Your
Resources
to Achieve:
• Sales Growth
• New Markots
• New Applications
• Product
Enhancements
• Profitable Pricing
Leslie C. Hruby, SM 73
F. Michael Hruby
326 Burroughs Rd.
Boxborough, MA 01719
(617) 263-0648
VenturesandAdventures
Start-Ups
Management
Consultation
Special Projects
Bridging
The Technical
and
The Organizational
The Bureaucratic
and
The Organizational
Specializing in the
solution of complete,
difficult, new,
Infrequent, or unique
problems
Giorgio Piccagll,
•67, Ph.D.
100 Dorado Terrace
San Francisco, CA
94112
(415)333-5084
probabilities of the condition occuring). I.e.,
p(Hl) = p(Hl:H2) x p(H2) + p(Hl:T2) x p(T2).
So
p(Hl) = (7/8)p(H2) + (1/2)|1 - p(H2)]
p(Hl) = 1/2 + <3/8)p(H2). (1)
Similarly, for Coin 2 we get
p(H2) = p(H2:Hl) x p{Hl) + p(H2:Tl) x p(Tl)
p(H2) = <7/8)p(H) + (1/2)[1 - p(Hl)J
p(H2) = 1/2 + (3/8)p(Hl). (2)
Combining equations (1) and (2),
p(Hl) = 1/2 + (3/8)[l/2 + (3/8)p(Hl)] = 4/5.
We also find that p(H2) = 4/5. Again using laws of
conditional probability, we know that
p(Hl & H2) = p(Hl:H2) x p(H2) = 7/8 x 4/5 =
Also solved by David DeLeeuw, Leon Tabak,
Matthew Fountain, Michael Jung, Richard Hess,
Winslow Hartford, and the proposer, William Stein.
FEB/MAR 3. A horizontal line of length 2a forms
the common base for two isosceles triangles. The
near side the triangle is 45° -45" -90°, and on the
opposite side 75° -75° -30°. Determine the radius
of the circle tangent to all sides of the composite
lanceolate figure, and locate the center.
Avi Omstein makes it look easy:
Let x be the radius of the inscribed circle, let h be
the line segment from the circle's center to the ver
tex of the 75° -75° -30° triangle, and let s be the
length of the bisector of this triangle. From the dia
gram, we see:
s = a/tan 15° and h - x/sin 15"
a + s = x21/2 + h.Thus we have
a + a/tan 15° = x21/2 + x/sin 15°
a(l + 1/tan 15^(2"* + 1/sin 15°)x = 0.896575472a
x2w = 1.267949192a
Also solved by David DeLeeuw, Everett Leroy,
George Parks, Harry Zaremba, Mary Lindenberg,
Matthew Fountain, Mel Garelick, Naomi Markov-
itz, Richard Hess, Steve Feldman, Winslow Hart
ford, and the proposor, Phelps Meaker.
FEB/MAR 4. Find a four-digit number whose square
is an eight-digit number whose middle four digits
are zero.
Most solutions were brute-force computer
searches. Pierre Hefller reduced the search vastly
by employing some pre-analysis:
The answer is 6,245, whose square is 39,000,025.
Trivial answers of 4,000, 5,000, 6,000, 7,000, 8,000
and 9,000 should have been excluded in the state
ment of the problem. Since the square must lie be
tween 10,000,001 and 99,000,099, the number itself
must lie between 3,163 and 10,000. Absent any way
to predict the occurrence of zeros in the middle of
a square, one could square each number between
3,163 and 10,000 (6,144 numbers in ail if endings in
zero are omitted) and hope to find a square with
four zeros in the middle. A tedious search without
a computer, almost two hours on my HP97. For a
more efficient search, consider the following: the
square root of 10,000,099 is 3,162.293; the square
root of 10,000,000 is 3,162.277. The difference is
0.0156. In the same test for numbers 99,000,099 and
99,000,000, the difference in square roots is 0.00498.
It follows that if X is a number ranging up from
10,000,000 and if VX comes out with a fractionwhich is more than 0.0156, there is no number be
tween X and X — 99 which is a perfect square. So,
using 0.0156 as a discriminant, take the square root
of 90 numbers of the form ab,0O0,O99, where ab
ranges from 10 to 99 (easy on any pocket calculator),
discard any square root if its fractional part exceeds
0.0156, and round out the rest to the next lowest
whole number. Then discard any left that end in
zero. Three remain that are worthy of being tested.
One of them, 6,245, satisfies the problem. The other
two do not because the discriminant did not de
crease to 0.00498 (as it should have to be a necessary
and sufficient test) as ab ranged up to 99. This search
took just over two minutes on my HP97.
Also solved by Jerry Cogan, Frank Carbin, Ches
ter Claff, Avi Omstein, Dennis Sandow, George
Byrd, George Parks, John Prussing, Lee Fox, Mat
thew Fountain, Michael Jung, Michael Tamada, Na
omi Markovitz, Nicholas Strauss, Richard Hess,
Robert Slater, Robert Turner, Ronald Raines,
Thomas Stowe, and Winslow Hartford.
FEB/MAR 5. Consider two dipoles. The lower di-
pole is fixed, and the upper dipole is constrained
to move along a horizontal line. (This is roughly the
geometry encountered in magnetic stirring.) Find
the conditions for which the upper dipole tends to
center (the force is in the opposite direction to the
displacement from the center line). When does the
motion of the upper dipole approximate simple har
monic motion?
Only Matthew Fountain and Richard Hess tac
kled this hard problem. Mr. Fountain's impressive
solution would have been published except for a
confrontation between its length and the available
space in this issue. Readers may obtain a copy (to
gether with our apologies, which go also to reader
Fountain) by return mail on request to the Review
office.
Better Late Than Never
JAN 4. Pierre Heftier has responded.
Proposers' Solutions lo Speed Problems
SD 1. 27.733
SD 2. First cash the *K, then lead the VQ. If this
holds (it did) and both defenders follow suit, play
the */J. Presumably, a defender will take the VK.
Now the ¥9 provides an entry to dummy, and you
can pitch the A10 on the ♦A. If East shows oul on
the first trump trick, play a small trump toward the
V9, with the same result. This hand occurred at
the Talleyville Club in Wilmington, Del., which
boasts of the highest percentage of life masters in
the U.S. West actually held three hearts to the king.
Several declarers made 12 tricks.
A24 JULY 198S
Goldbetg-Zaino&Associates
Geotochnlcal-
Goohydrologleal
Consultants
Foundation/Soil
Engineering
Site Planning &
Development
Soil/Rock Testing
Geotechntcal
Instrumentation
Testing of Construction
Engineering for
Lateral Supportftintrnmi
Rock Engineering
Grcundweter
Engineering
Underground
Construction
Inc
D. T. Goldberg, '54
W. S. Zolno, '54
J. D. Guertln, '67
M. J. Barvenlk, 76
M. 0. Bucknom, '81
N. A. Campagna, Jr. '67
F. W. Clark, 79
W. E. Hsdga, 79
W. E. Jaworskl, '73
C. A. Undberg, 78
R. M. Simon, 72
E. 1. Steinberg, '80
T. vonRosenvlnge (V, 'BO
The GEO Building
320 Needham St
Newton Upper
Falls, MA 02)64
(617) 467-8640
EdwaidR
Maiden
Gup.
Builders tor Industry,
Institutions, hospitals,
manufacturing plants.Qovsriwwnt srtd
developers of high
technology facilities o
for over 35 years
Edward R. Marden '41
Douglas R. Marden '62
280 Lincoln Street
Boston, MA 02134
(617) 782-3743
JamesGoldstein&Partners
ARCHITECTS
ENGINEERSPI AMKIPRQ
S. Jamas Goldstein '46
Etlot W. Goldstein '77
225 Mlllbum Avenue
Mlllburn, NJ 07041
(20t) 467-8840
SERVING THE SCIENCE COMMUNITY SINCE 1953
FACILITIESResearch 4
Development
Education a Training
Management &
Support
Manufacturing &
Warehousing
FOR HIGH
TECHNOLOGYFIELDS
Biochemistry
Chemical Engineering
Chemistry
Computer Science
Electronics
EMI Structures
Hazardous Materials
Information Science
Laboratory Animals
Lasers
Medical Sclencos
MlcroelectronIcs
Monoclonal
Antibodies
Particle Accelerators
Pulp ft Paper
Recomblnant DNA
Solid State Physics
Telecommunications
Toxicology
Wind Tunnels
PUZZLE CORNER
ALLAN J. GOTTLIEB
Fun with Trig
Since it has been over a year since I
reviewed the criteria used to select
solutions for publication, let me do
so now.
As responses to problems arrive, they
are simply put together in neat piles,
with no regard to their date of arrival or
postmark. When it is time for me to
write the column in which solutions are
to appear, I first weed out erroneous and
illegible responses. For difficult prob
lems, this may be enough; the most pub-
lishable solution becomes obvious.
Usually, however, many responses still
remain. I next try to select a solution that
supplies an appropriate amount of detail
and that includes a minimal number of
characters that are hard to set in type.
A particuarly elegant solution is, of
course, preferred. I favor contributions
from correspondents whose solutions
have not previously appeared, as well
as solutions that are neatly written or
typed, since the latter produce fewer
typesetting errors.
Finally, let me credit David Griese-
dieck as the solver of OCT 3. Somehow
his name disappeared when the column
travelled from New York to Cambridge.
Problems
M/Jl. We begiii with a two-part bridge
problem from Doug Van Patter based on
tho following deal:
A K
VQJ
♦ Q9
+ Q
♦
♦
*
9432
832
A
¥
*
A84
AK10
A765
KJ
9765
7
K10
A986
4
♦ QJ¥86
♦ J+ 10/
32
10 3 2
5
'54
How does South make five clubs after
an opening lead of the ♦ Q, and what
lead sets this contract?
M/J 2. As I look out my kitchen window
and see our lake frozen solid, I am
warmed by Phelps Meaker's N-sided
flower garden described in the following
problem, entitled "Fun with Trig":
A large bed of flowers and greenery
is laid out in the form of a regular pol
ygon of N sides. A walk composed of N
trapezoidal concrete slabs surrounds the
flower bed. A circumscribing circle pass
ing through the outer corners of the
walk and an inscribing circle tangent to
the inner flats of the trapezoids have cir
cumferences in the ratio of almost ex
actly 191:165. The total area within the
outer periphery of the walk and the area
of the walk itself are in the ratio of 4:1.
Find N.
M/J 3. Lester Steffens wants you to ex
plain a parlor trick he is fond of:
Given a number N between 500 and
1000, rapidly construct a series of num
bers using eight of the nine digits from
1 through 9 once, the other digit twice,
and a few zeros so that the series totals
N. Example: if N = 642, a solution is
10
20
334
50
60
78
_90
642
M/J 4. Winslow Hartford asks a mniii-
faceted problem:
As we generate geometric figures to
represent y = x", we have "elements"
consisting of points, lines, faces, cubes,
etc. as n increases. For the number of
points in each figure, we have (for n >
0) P = 2". Derive the number of lines
and faces in a five-dimensional hyper-
cube.
M/J 5. Eric Schonblom tells us about his
8.01 doodles, with an apology for tar
diness: "Having discovered this during
SEND PROBLEMS. SOLUTIONS,
AND COMMENTS TO ALLAN I.
GOTTLIEB. -67. ASSOCIATE
RESEARCH PROFESSOR AT
THE COURANT INSTITUTE OF
MATHEMATICAL SCIENCES.
NEW YORK UNIVERSnr. 251
MERCER ST.. NEW YORK.
N.Y.. 10012.
A22 MAY/JUNE 198.5
OBITUARIES
Highway Safety Pioneer
Lived a Life "Deserving of Prominence
A TRIBUTE BY RALPH NADER
William Haddon Jr.
1926-1985
Physician William Haddon, Jr., '49, a leader
in highway accident research, died in Wash
ington, D.C. on March 4 of the complica
tions of kidney failure. Consumer advocate
Ralph Nader believes that Dr. Haddon's ca
reer owed much to his M.l.T. education and
that his death at 58 was a tremendous loss
to the cause of highway safety. In many
ways, Dr. Haddon's life exemplified the val
ues propounded by Nader when he gave the
keynote address at the Alternate Jobs Fair,
organized in late winter by the M.l.T. Stu
dent Pugwash (see facing page).
William Haddon, Jr.
"T oss prevention" is not a conven-
I tional focus for engineers and phy-J—Jsicians, but Dr. William Haddonperformed so brilliantly in that field that
students aspiring to technical and med
ical careers can benefit greatly from an
awareness of his contributions.
After obtaining degrees from M.l.T.,
Harvard Medical School, and the Har
vard School of Public Health, Haddon
spent 10 years with the New York State
Department of Public Health. He was
concerned with highway traffic casual
ties, which at that time attracted few
minds possessing staying analytic
power and follow-through stamina,
traits Haddon had in ample supply.
Motor vehicle crashes and trauma
were failures in energy management, in
Haddon's view. He separated the ve
hicular crash phenomenon into pre-
crash, crash, and post-crash stages in
order to clarify the intervention appro
priate for each stage. Haddon demon
strated how the engineering
intervention known as crashworthi-
ness—exemplified by air bags—could
manage the lethal transfers of energy in
a crash to save life and limb. Through
his studies on alcohol-related crashes,
he was familiar with those data. But he
believed that the engineering approach
had a greater chance of reducing death
and injury than any available method for
controlling the complex behavioral
sources of drinking and driving.
This emphasis on engineering reme
dies was rooted in Haddon's knowledge
of empirical successes throughout his
tory (which he cited frequently)—from
the Greek physician Hippocrates to ad
vances in factory and railroad safety for
workers—and in his detailed knowledge
of crash protection systems which were
available but not in use.
In the 1960s, Haddon was part of the
growing consumer challenge to the au
tomobile industry though his writings,
testimony, and television and radio in
terviews. From 1966-1969, he was the
first administrator of the National High
way Traffic Safety Agency established
by President Johnson. Under Haddon's
direction, the agency issued motor ve
hicle safety standards, conducted re
search, and required the recall of
defective vehicles and tires.
After leaving government, Haddon
became the president of the Insurance
Institute for Highway Safety in Wash
ington, where he took the lead in trans
forming the insurance industry into a
critic of unsafe and costly vehicular en
gineering. He also helped establish the
Highway Loss Data Institute, which
provided safety information on vehicles
by make and model.
The words that come to mind when
reflecting on Haddon's work are rigor
ous, dedicated, unyielding in purpose,
humane, blunt, and a long-distance run
ner. He traversed the whole contin
uum—making major contributions to a
conceptual foundation for trauma pre
vention, to public education, safety reg
ulation and research, and to a new
perception of the mission of the insur
ance industry. Haddon also raised the
status of professional work in trauma
prevention—introducing younger mem
bers of relevant professions such as sta
tistics, engineering, and biology to the
cause of saving lives through knowledge
and its application.
Haddon's was a life to be celebrated,
deserving of prominence, in the hope
that many will follow in his footsteps if
they only know of his trail. □
DeceasedThe following deaths have been reported to theAlumni Association since the Review's last deadline:
John B. Welch, '13; October 3, 1984; Indianapolis,Ind.
Atwood P. Dunham, '17; December 14, 1984; Ded-ham. Mass.
Raymond E. McDonald, '17; October 18, 1984; Na-lick. Mass.
Freeman H. Dyke, '20; March 8, 1984; Tequesta,Fla.
Fraser M. Moffat, '20; October 24, 1984; Montrose,Penn.
William D. Shepard, '20; January 1985; Winnetka,
James J. Wolfson, '20; November 14, 1984; Hallan-
dale, Fla.
Chester A. Rimmer, '21; February 7, 1985; Norwell,Mass.
George T. Boli, '22; 1985; Venice, Fla.
John F. Hennessy, '22; February 9, 1985; ChestnutHill, Mass.
W. Raymond Hewes, '22; December 24,1984; Need-ham, Mass.
James D. Sarros, '22; December 27, 1984; Madison,N.J.
Mrs. Francis W. Spalding, '22; July 1984; Cincinnati, Ohio.
Arthur R. Belyea, '23; December 16, 1984; Old Say-
brook, Conn.
Arthur Raymond Holden, '23; February 11, 1985;
Sarasota, Fla.
PHOTO: INSURANCE INSTITUTE FOR HIGHWAY SAFETY TECHNOLOGY REVIEW A21
a physics lecture over 30 years ago, I'm
a little slow in sharing it. It's a paper-
and-pencil puzzle but is most easily
stated in terms of scrabble tiles":
Take the letters in the first half of the
alphabet in order. Place the A on the
table. Place the B next to one of the four
sides of the A. Place the C next to the
one of the six sides at the AB (or BA)
pair. Then add the D and so on. If you
do this correctly, when you reach the
letter M you will have created a cross
word-puzzle matrix of complete com
mon English words, no proper names,
no foreign words, and no acronyms or
abbreviations. Having solved the prob
lem as posed, can you add one or more
letters to the A-M set and still retain
complete words?
Speed Department
SD 1. Jerry Grossman wants to know
what is so interesting about
f(x) = x1***,
where log is the natural logarithm.
SD 2. David Evans designed a single
elimination tennis tournament for 37
contestants to have the minimum num
ber of byes. How many matches were
played?
Solutions
JAN 1. South is on lead and is to take six tricks
against best defense with hearts as trump:
A 9832
V 10
♦ AQ
* —
J¥ 8
♦ ) 10 9 8
♦ 5
* 10
V 6
♦ K6
AKJ6*AK7y
♦ —
* AQ4 3
Ben Feinswog solved the problem by losing the
A7 and discarding the * A and *K:
South plays the *A (removing West's exit), dis
carding the ♦ Q from dummy, and then leads the
*7 to West's A). On the forced red-card return.
South plays dummy's winning ¥10, and ♦ A, dis
carding the *A and 4>K from hand, and claims the
balance with dummy's last three spades.
Also solved by Avi Ornslcin, Doug Van Patter,
Edgar Rose, Ellen Kranzer, Joe Hahn, John Lacy,
Larry Wischhoefer, Matthew Fountain, Red Clev-
enger, Robert Lax, Roy Schweiker, Tim Maloney,
Walter Cluett, Winslow Hartford, Jim Landau, John
Rule, Richard Hess and Emmet J. Duffy.
JAN Z. A smooth, rigid, and circular hoop hangs
from a rigid support by an ideal, extensionless
string. Two small beads slide along the hoop (like
beads of a necklace) with negligible drag and fric
tion. The beads are slid to the top of the hoop and
released. How massive must each bead by lo spon
taneously lift the hoop?
Matthew Fountain was pleased to submit a so
lution but even more pleased to report that his wife
has responded perfectly to a recent cataract oper
ation:
The two beads must total three rimes the mass of
the hoop. As each bead slides outward, its circular
motion causes a centrifugal force with an upward
component. At the same time its downward accel
eration increases, decreasing the force that it exerts
upon the hoop. The maximum lifting force occurs
when the sum of these two effects is greatest. Until
the hoop actually moves, a bead does no work.
Therefore, its gain in kinetic energy equals its loss
in potential energy. Thus, (l/2)mvy =■ smg, wheres = vertical drop, m - mass, g - gravitational
constant, and v = velocity.
\When the bead has traversed an arc <(> on the hoop
of radius r, the vertical drop
s = r(l - cos *).
The centrifugal force is mv*/r, with an upward component
(mvVr)cos 4> = 2smg(cos <l>)/r = 2mg(l -cos <|>) cos
<t> . Gravity produces a force mg acting downward
through the bead. When the bead has traversed
through the arc 4>, the component of this force
toward the center of the hoop is mg(cos<t>). In rum,
the downward component of the radial component
is mg(cos! <)>). The lifting force F exerted by a beadis
F = 2mg(l — cos <t>)cos <J> — mg(cosJ 6) = 2mg(cos
<f>) - 3mg(cos2 *).
The maximum and minimum values of F occur
when
dF/d<J> = - 2mg(sin 4>) + 6mg(cos <J>)(sin 4) = 0.
Thus the minimum lift occurs when sin 6 = 0 and
the maximum lift occurs when cos <S> = 1/3:
F™, = (l/3)2mg - (1/3)2 3mg = (l/3)mg.
Each bead will lift up to one-third of its own weight.
Also solved by David Smith, Gary Heiligman,
Harry Garber, Harry Zaremba, John Lacy, John
Prussing, Ken Haruta, Jim Landau, Richard Hess,
Peter Kramer, and the proposer, Bruce Calder.
JAN 3. A man received a check calling for a certain
amount of money in dollars and cents. When he
went to cash the check, the teller made a mistake
and paid him the amount which was written in
cents in dollars, and vice-versa. Later, after spend
ing S3.50, the man suddenly realized that he had
twice the amount of money the checked called for.
What was the amount on the check.
Red Clevenger provides a "down home" solu
tion:
The problem reminds me of my eighth-grade
teacher in rural Afton, Okla., who had several dol-
lar-and-cents algebraic problems which she always
wanted us to solve by using one variable. I preferred
using two variables in simultaneous equations. She
did drive home the lesson, however, of lOOtf = 15.
Let
$ = the dollar amount of the check, and
c = the cents amount;
then the problem stated algebraically is:
lOOe + $ - 350 = 2(100$ + c). (1)
The other equation results from the difference be
tween the number of cents received (S) and twice
the number of cents on the check (c) which must
be - 50 since c is greater than S. Staled algebraically:
S - 2c = - 50. (2)
Solving (1) and (2) results in S = 14 and c - 32,
and thus the check amount was SI4.32.
Eloctricsl contractors
to the nation alneo
1895
Headquarters:
45 Rockefstlar Plaza
Now York, N.V. 10111
Offices In 16 principal
cltlss throughout tho
U.S. and Puerto Rico
Boston Office:
86 Coolldge Avo.
Watertown, MA 02173
(617) 926-5500
Alexander W. Moffat, Jr.
HE
Hawkins
&SonsCompany
Building contractors
Steven H. Hswklnt, '57
168 Whfflng Street
Kingdom, MA 02043
(617)749-6011
(617) 7494012
DebesQnp.
Health Care Consultant*
Design, Construction,
Management
Subsidiaries:
Charles N. Debes A
Assoclnc.
Alma Nelson Manor Inc.
Pork Strathmoor
Cofporotloo
RocMord Convalescent
Center Inc.
Chambro Corporation
Charles N. Debes '35
5663 Strathmore Drive
RocMord, ll_ 61107
TECHNOLOGY REVIEW A23
VenturesandAdventures
Start-Ups
Management
Consultation
Special Pro|octt
Bridging
Tho Technical
and
The Organizational
Tha Bureaucratic
and
Ths Organizational
Spsclallzlng In tha
solution of complete,
difficult, now,
Intrsquont, or unique
problems
Qtorglo Pfccao.II,
'67, Ph.D.
tOO Oerado Torrace
San Francisco, CA94112
(415)333-5084
Gecnge A.Roman &Associates Inc
Architecture Planning
Interior Doslgn
Institutional
Commercial
Industrial
Reeldentlal
Slto Evaluation
Land Use Planning
Matter Planning
Programming
Intorlor Space
Planning
Colleges
Hospitals
Medical Building!
Office Buildings
Apartments
Condominiums
Qsoroo A, Roman,
A.I.A.'65
One Gateway Center
Newton, MA 02158
(617)332-5427
Consulting Geotechnlcal
Engineers and
Geologists
Soil and Rock
Mocnomcs
Engineering Geology
Engineering Geophysics
Foundation Engineering
Terrain Evaluation
Engineering Seismology
Earthquaka Engineering
Geohydrology
238 Main St.
Cambridge, MA 02142
(617)492-6460
Horl P. Aldrich, Jr. '47
Martin C. Murphy 'SI
Edward B. Ktnner '67
Douglas G. Grrford 71
Joseph J. Rixner '68
John P. Ougan '68
Kenneth L Rocker '73
Mark X. Haley '75
Robin B. Dill '77
Andrew F. McKown 78
Keith E. Johnson 'SO
Also solved by Winslow Hartford, Allan Benson,
Anthony Lombardo, Avi Ornstcin, Charles Sutlon,
David Smith, Edgar Rose, Frank Carbin, Fred Sleig-
man, Gary Drlik, George Byrd, George Aronson,
Harry Garbcr, Harry Zarcmba, Howard Stern,
James Michelman, John Lacy, John Prussing, Ken
Haruta, Larry Wischhoefer, Leon Tabak, Marion
Berger, Matthew Fountain, Michael Jung, Naomi
Markovitz, Norman Spencer, Peter Silverberg,
Phelps Meaker, Ronald Martin, Steve Feldman, Ted
Numata, Thomas Stowe, Everett Leroy, Tim Ma-
loney, Jim Landau, Ronald Raines. Richard I less,
George Parks, and the proposer John Rule.
JAN 4. A rigid arm pivots around the fixed point
A. At the end of the arm is a follower (B) which
runs in a curved track. The track pivots about the
fixed point C. If AB = AC = r, find the shape of
the track such that its slope at C is always vertical.
The problem as stated was rather easy. Surpris
ingly, several readers noticed that the diagram in
dicated a vertical slope at B, not C. This was indeed
more interesting. Charles Sutton writes:
The problem as stated is trivial, since clearly a track
in the shape of a semi-circle of radius r with center
M.I.T. ALUMNI
CAREER SERVICES
GazetteA listing every two weeks
of jobs for alumni across
the country
We want more firms to
know they can list jobs in
the Gazette
We want more alumni mak
ing a job change to know
the Gazette can direct them
to job opportunities
Whether you have a job to
fill, or are looking for a job,
let us send you a copy of
the Gazette to show you
how it can help you
Call or write
Marianne Ciarlo
Alumni Career Services
M.I.T., Room 12-170
Cambridge, Mass 02139
Tel: (617) 253-4735
at A would remain stationary and its slope at C
would always be vertical. I assume what was in
tended was that the slope of the track at B should
always be vertical. This really had me going for a
while. I set it up in rectangular coordinates, using
anlaytic geometry and calculus, and ended up with
an impossible differential equation. Then [ tried po
lar coordinates, used a few trig identities and cal
culus, and got a real simple differential equation
that integrated to give me a circle. Once you know
what you're looking for, you need only elementary
geometry. For the slope at B to remain vertical, the
shape of the curved track is determined by the facts
that AC = AB = r and the tangent to the track at
B is perpendicular to the line at AC. Imagine that
the track is held fixed; then point A will have to
move on a circle of radius r. The accompanying
diagram shows that the track is an arc of a circle of
radius r with center at D. The tangent to the circle
at B is perpendicular to both BD and AC, so AC
must be parallel to BD. Also AC = BD - r, so ABCD
is a parallelogram and AB = CD - r.
Also solved by David Smith, Gary Heiligman,
George Byrd, Harry Garber, Harry Zaremba. How
ard Stem, John Lacy, Jordan Wouk, Matthew Foun
tain, Phelps Meaker, Red Clevenger, Winslow Hart
ford, Richard Hess, Peter Kramer, and the pro
poser, Floyd Kosch.
Better Late Than Never
Y19M Randall Whitman, William Thompson, Rick
Lufkin, Marion Berger, Al Weiss, Matthew Foun
tain, Donald Trumpler, and Alan Katzenstein each
improved on the published solution. When com
bined, their efforts yield the following list of revi
sions:
26 = (1 + 9/4) x 8
31 = 49 - 18
37 = 1 + 9(8 - 4)
41 = 49 - 8 x 1
42 = (49 - 8) + 1
49 x Is
81 - (4 + 9)
V x 8/4
4-1"
1" x 4
1" + 4
(49 - l)/8
91-84
8 x 1«
= 1" + 914 = 14 x (9 - 8)
16 = (1 + 9 - 8)4
18 = 9 x 4 - 1822 = (89 - l)/4
49 =
68 =
70 =
76 =
81 =
99 =
100
(9 - 1/4) x 8
94-18
4 x 18 + 9
98 + I4
= (1 + 9**
JUL 2. Howard Stern notes that this problem ap
pears in Martin Gardner's Mathematical Circus.
OCT 3. John Stackpole notes that the very week his
F/M 1 issue of Technology Review arrived, the Mary
land Lotto game had ten participants match all six
numbers and 827 match five out of six, giving
strength to Griesedieck's assertion that the selection
of numbers by participants is not random. Jonathan
Hardis, the proposer, remarks that never have any
of the large lotteries had a positive expected return
for the player and also expresses the oft-stated po
litical belief that these games constitute a regressive
tax since they played to a disproportionately large
extent by the poor.
Proposers' Solutions lo Speed Problems
SD 1. It is constant, i.e. independent of x.
SD 2. 36 One contestant is eliminated each match.
A24 MAY/JUNE 1985
PUZZLE CORNER
ALLAN J. GOTTLIEB
Can You Find the Bad Penny in the Bank?
Just as I am writing this the mailman has
brought a letter from the New York State
Crime Laboratory marked "official busi
ness." After two gulps and one re
flection as to what I could have done that
they found out about, I got up my cour
age to open the envelope and was re
lieved to find it was just a "Puzzle
Corner" reader responding to several
problems from the January issue.
As I have remarked previously, Nob-
uyuki Yoshigahara selected me as
"World Puzzlist" No. 8, and now he has
forwarded the issue of Quark in which
this honor was officially bestowed. In
addition to being flattered, I enjoyed
hearing the transliteration of "Gottlieb"
when one of my Japanese-speaking col
leagues read the beginning of the col
umn. Thank you again, Mr.
Yoshigahara. I should also mention that
long-time "Puzzle Corner" contributor,
Richard Hess, was selected World
Puzzlist No. 6.
Problems
APR 1. I read our first problem for this
month in net.chess, an electronic news
group devoted to chess. Roughly speak
ing, these newsgroups consist of widely
separated individuals who communi
cate with each other via electronic mail.
I especially enjoyed the following two-
part offering from Jeffrey Mattox, who
noted that it is possible, albeit unlikely,
for the position to occur in a game:
White is to play and mate in two. Mattox
notes that at first glance there appears
to be two possible solutions. You are to
show that only one meets the need.
APR 2. Our next problem is from Phelps
Meaker, who first asks you to study:
16601.92 + 14374.08 =
11334.4 + 19641.6 =
18521.44 + 12454.56 =
4147.36 + 26828.64 =
He then notes that each pair is a differ
ent way of evaluating the same equation
and asks you to write the equation in
the usual form. He also offers a hint, but
you may wish to try the problem with
out this aid. [The hint is to note which three
numbers are perfect squares.]
APR 3. Rich Decker wants you to find
all maxima and minima of
without using any calculus. This problem
appeared in an Ohio State University
prize exam for undergraduates.
APR 4. Allan Faller wants us to be penny
wise and writes:
On each day of the year (not leap year)
you are given a penny. On December 31
you are given your last penny and told
that it was fresh from the U.S. Mint, but
that one of the previous pennies may
have been counterfeit, and therefore
lighter or heavier. than the standard
penny. You are asked to determine the
number of balancings, using a common
pan balance, that would be necessary
and sufficient to determine whether or
not there is a counterfeit coin, and if
there is, to tell whether it is heavier oflighter than the last penny that you re
ceived.
APR 5. Our final regular problem, from
Martin Brock, is based on the familiar
"crossed ladders" configuration at the
bottom of the previous column:
Mr. Block asks you to find X for two
configurations. First when A = 15, B =
10, and H = 8; and second when A =
16 + 2V2, B = 16 - 2V2, and H = 2.
Speed Department
SD I. A bridge quickiePaftor*1 atier.
North:
A K854
¥ Q 10 7 4
♦ 53
♦ AQ 10
South:
♦ A93
¥—
♦ AQ6
*K987643
East: South:
IS 2C
3H 5C
5H 6C
West
P
P
D
from Doug Van
: North:
3C
P
P
Instead of defending the usual five-heart
bid by East, you (South) make the ag
gressive bid of six clubs. West opens
with ¥5, which draws the ¥10, ¥K,
and a trump. You lead a club to dum
my's «fcA and East shows out. Your
finesse of the *Q loses to West's 4K,
and West returns a trump to the *10.
You lead to the 4A (East shows out),
and ruff your third diamond with dum
my's last club. Can you find a way to
justify your overbid? (East is an excellent
player, never known to psych).
SD 2. Joseph Horton writes: Great news!
I have answered an age-old question:
Which came first—chicken or egg?
SEND PROBLEMS, SOLUTIONS.
AND COMMENTS TO ALLAN /.
GOTTLIEB. 67. ASSOCIATE
RESEARCH PROFESSOR AT
THE COURANT INSTITUTE OF
MATHEMATICAL SCIENCES.
NEW YORK UNIVERSITY. 251
MERCER ST.. NEW YORK.
N.Y.. 10012.
A.)0 APRIL 1985
■ OBITUARIES,
Jerome H. Holland,
1916-1985
Jerome H. Holland, prominent black
educator and civil rights advocate
who was a member of the Corpora
tion from 1969 to 1979, died in New
York on January 13 of cancer; he was 69.
Dr. Holland studied sociology at Cor
nell and the University of Pennsylvania
and went on to become president of two
predominently black colleges, Delaware
State College and later Hampton Insti
tute, before accepting appointment as
Ambassador to Sweden.
In addition to his service on the M.I.T.
Corporation and its visiting committees
for the Center for International Studies,
Department of Humanities, and ath
letics and student affairs. Dr. Holland
was national chairman of the American
Red Cross and the first black on the
board of the New York Stock Exchange.
Deceased
Robert S. Beard, 05; 1983; Trinidad, Calif.
Mrs. Milton E. Hayman, 11; 1984; West Hart
ford, Conn.
Clifford L. Muzzey, Sr., '14; December 15, 1984;
Sandusky, Ohio.
Solomon Schneider, '15; October 12, 1984; Haver-
town. Penn.
James B. Hobbs, '16; January 6, 1985;. Natick,
Mass.
Donald B. Webster, 16; November 25, 1984; Fal-
mouth, Mass.
Benjamin I. Lewis, '17; December 21, 1984; Ta-
coma, Wash.
R. Parry Kennard, '18; 1984; New York, N.Y.
John S. Coldwcll, 19; September 21, 1984; Fall
River, Mass.
Shee M. Lee, 19; 1984; Hsinchu, Taiwan.
Chester C. Stewart, 19; May 18, 1984; Needham.
Mass.
George H. Wiswall, Jr., 19; 1982; Edgartown.
Mass.
Preston W. Smith, '21; January 6, 1985; North
Weymouth, Mass.
Charles Kerr, Jr., '22; Fort Point, Fla
Dwight E. Stagg, '22; November 13, 1984; Bridge
port, Conn.
William P. Winsor, '23; December 16, 1984; New
York, N.Y.
Cordon H. Crabb, '24; August 1, 1984; Winter
Park, Fla.
Mrs. Theodore G. Coyle, '25; July 12, 1984; Nor
wood, Mass.
Marvin H. Green, '25; November 6, 1984; N.
Palm Beach, Fla.
Eugene C. Hermann, '25; August 20, 1984; West-
field, N.J.
George Oetinger, Jr., '25; 1984; Monticello, Fla.
Peter H. Sin, '25; September 13, 1984; Hong
Kong.
Mrs. George V. Slottman, '25; January 5, 1984;
New York, N.Y.
Elmer C. Warren, '26; October 11, 1984, Waler-
ville, Maine.
Adelbert N. Billings, '27; November 13, 1984;
Richmond, Va.
Laurence H. Coffin, '27; November 11, 1984;
North Conway, N.H.
George D. Fexy, '27; September 1984; Kirkland,
Wash.
Randolph J. Peterson, '27; November 4, 1984;
Rochester, N.Y.
John M. Ryan, '28; November 3, 1984; Winches
ter, Mass.
John D. McCaskey, '29; January 10, 1985; St. Jo
seph, Mo.
Theodore Criley, Jr., '30; January 1985; Clare-
monl, Calif.
Norman H. Dolloff, '30; October 3, 1984.
Harry J. Fekas, '30; November 2, 1984; Newport
News, Va.
Emile P. Grenier, '31; December 12, 1984; Ann
Arbor, Mich.
John Vasla, '31; 1984; Punta Gorda, Fla.
George E. Colby, '32; October 3, 1984; Westport
Harbor, Mass.
John F. Crowther, '32; October 19, 1984; Old
Lyme, Conn.
George Hcnning, '33; November 21, 1984; Syos-
set, N.Y.
Walter S. Brodie, '34; January 20, 1985; Marble-
head, Mass.
Karl A. Gardner, '34; January 3, 1985; North-
ridge, Calif.
Herbert L. Gamer, '34; November 26, 1984; Mil-
Ion. Mass.
Proctor Wetherill, '34; January 16, 1985; Chester
Springs, Penn.
Stephen H. Richardson, '36; 1984; Seattle, Wash.
Robert M. Sherman, Jr., '36; November 27, 1984;
Warwick, R.l.
Staunton L. Brown, '38; October 8, 1984; Madi
son, Conn.
Roscoe J. Cooper, '38; August 27, 1984; Beverly,
Mass.
Edgar H. Kiblcr, Jr., '39; 1983.
Michael Morelli, '40; November 16, 1984; Alexan
dria, Va.
George Farneil, '41, September 2, 1984; Syracuse,
N.Y.
J. Nelson Evoy, Jr., '42; November 1984; Bryn
Mawr, Penn.
Howard W. Comey, '43; 1984; South Acton,
Mass.
Howard Weaver, Jr., '44; 1983; Phoenix, Ariz.
Edmond J. McClure, Jr., '45; 1983; Blasdell, NY
James W. Shearer, '45; October 9, 1984; Liver-
more, Calif.
Charles H. Tavener, Jr., '45; Jauary 17, 1982, Boca
Raton, Fla.
James W. Church, '46; 1983; Silver Spring, Md.
Norman P. Hobbs, '48; October 17, 1984; Chest
nut Hill, Mass.
George H. Bradley, Jr., '49; April 19, 1984; Albu
querque, N.M.
Wilmer S. Garwood, Jr., '50; December 8, 19S4
Ben Silver, '50; 1982, Barrington, R.I.
James B. Law, '52; 1984; New Delhi, India.
Vernon V. Hukee, '56; September 26, 1984; Na
shua, N.H.
William L. White, '59; January 21, 1985; Newton-
ville. Mass.
Mohammad R. Damghany, '62; December 22,
1984; Arlington, Mass.
William C. Margetts, '62; 1984; Cambridge, Mass.
Edward Terner, '65; September 27, 1984; Shel-
bume, Vl.
Louis J. Urban, '70; October 31, 1984; Kettering,
Ohio.
James S. Steil, '80; December 20, 1984; Moses
Lake. Wash.
M.I.T. ALUMNI
CAREER SERVICES
GazetteA listing every two weeks
of jobs for alumni across
the country
We want more firms to
know they can list jobs in
the Gazette
We want more alumni mak
ing a job change to know
the Gazette can direct them
to job opportunities
Whether you have a job to
fill, or are looking for a job,
let us send you a copy of
the Gazette to show you
how It can help you
Call or write
Marianne Ciarlo
Alumni Career Services
M.I.T., Room 12-170
Cambridge, Mass 02139
Tel: (617)253-4735
TECHNOLOGY REVIEW A29
Solutions
N/D 1. While, moving first, is to mate in threemoves:
Howard Stem found this problem to his liking:
The following three plays represent the only possibilities resulting from While's initial move:
While:
Q-C2Kn-d3
Q-b2 mate
While:
Q-c2
Kn-f3
Q-d2 mate
While:
Q-C2
Kn-d3
Q-a2 mate
Black:
KM
K-a3
Black:
K-dA
K-e3
mack:
K-b4
Also solved by Eric Rayboy, David DeLeeuw,
R. Bart, Benjamin Rouben, Matthew Fountain,
Steve Feldman, Ronald Raines, Elliott Roberts,
. and the proposer, J. Weatherlv.
N/D 2. Find a number thai equals its own logarithm.
There is no positive real number x such thateither logl0(x) = x or log^x) = x. Several readers,
including the proposer. Smith D. Turner (/dt),
went on to consider log,. However. I do not feel
this meets the conditions of the problem (note
that logarithm was used in the singular). Tim Ma-
loney (and others) made another generalization;he writes:
First, one must recognize that the solution is a
complete number, call it Z = u + iw = re"Then
re* = ln(re*)
r-eosO + irsin 0 = In (r) + i 6, or
ln(rVr = cos 8
r = 0/sinO.
We must therefore solve
In(6/sin8) » 8-cot fl.
An iterative solution gives
8 = 1.337236 .... so
Z = (1.374557 . . .)e": 5)r2* '
= (.3181313 . . .) + 1(1.337236 . .).
I must admit that I first saw this problem in my
junir.: year at M.l.T. (1970), when someone pro-
post ' the problem in a lunchtime discussion in
Protessor Daniel Kleppner's research group.
Klenpner immediately drew a graph on the board
to ove the solution could not be real, asserted
that it must be complex, and left us all speech
less.
Also solved by Eric Rayboy, R. Bart, Matthew
Fountain, Ronald Raines, Winslow Hartford, John
Spalding, John Woolston, Naomi Markovitz, Mike
DebesCaip.
Haattti Caro Consultants
Design, Construction,
Management
Subsidiaries:
Charles N. Debes &
Assoc. Inc.
Alma Nelson Manor Inc.
Park Strathmoor
Corporation
RocMord Convatoscent
Center Inc.
Chambro Corporation
Charles N. Debes '35
5668 Strathmoro Drive
RocMord. IL. 61107
Goldbetg-Zoinn&Associates line
Geotochnlcal-Geohydrologlcal
Consultants
Foundation/Soli
Engineering
Site Planning &
Development
Soil/Rock Testing
Gootechnlcal
Instrumentation
Testing ol Construction
Engineering lor
Lateral Support
Systems
Rock Engineering
Groundwater
Engineering
Underground
Construction
D. T. Goldberg, 'S4
W. S. Zolno, '54
J. O. Guortln, '67
M. J. Barvonlk, 76
M. 0. Bucknam, '31
N. A. Campagna. Jr. '67
F. W. Clark, '79
W. E. Hodge, '79
W. E. Jaworskl, 73
C. A. Undberg, 78
R. M. Simon, 72
E. I. Steinberg, '80
T. vonRosonvingo IV, '80
The QEO Building
320 Needham St.
Newton Upper
Falls, MA 02164
(617) 467-8840
PaulEDutelle&GompanyInc
Roofers and Metal
Craftsmen
P.O. Box 96
Newtonvlllo, MA 02160
Hawkins&SonsCompany
Building contractors
Sloven H. Hawkins, '57
188 Whiting Street
Hinghem, MA 02043
(617) 7494011
(617) 749-6012
IheGodmanGampany,Inc
Industrial and
Commercial Real Estato
Mark Gottosman 70
(M.C.P.)
211 Congress Stroot
Boston, MA 02110
(617) 4234500
James
Goldsfpin&
Partners
ARCHITECTSENGINEERSQ| AMMCDCrLANNcno
S. James Goldstein '46
Eliot W. Goldstein 77
225 Mlllbum Avenue
Millburn, NJ 07041
(201) 467-6840
SERVING THE SCIENCE COMMUNITY ONCE 1133
FACILITIESResearch &
Development
Education & Training
Management &
Support
Manufacturing »Warehousing
FOR HIGHTECHNOLOGY
FIELDS
Biochemistry
Chemical Engineering
Chemistry
Computer Science
Electronics
EMI Structures
Hazardous Materials
Information Sclenco
Laboratory Animals
Lasers
Medlcel Sciences
Microelectronics
Monoclofiw
Antibodies
Particle Accelerators
Pulp & Paper
Recomblnant DNA
Solid State Physics
Telecommunications
Toxicology
Wind Tunnels
TECHNOLOGY REVIEW A31
ThomasKDyer,
Inc.
DIVISION OF HNTB
Consulting Engineers
Rail Transportation
Thomas K. Dyor '43
1762 Massachusetts Ave.
Lexington, MA 02173
(617) 862-2075
Washington, D.C.
(202) 466-7755
Chicago, IL
(312) 663-1575
Philadelphia, PA
(215) 569-1795
Nelson,Goulson&
Associates,Inc
Professional Suiting
Consultants
Contract Engineering
Services
Professional
Recruiting Services
Technical Personnel in
All Disciplines
333 W. Hampdan Ave.
Suits 507
Englewood, CO SOU 0
(303) 761-7680
Other offices In
Albuquerque,
Colorado Springs,
Dallas & Soattle
Paul R. Coulson
PE,'43
President
Edward R.MaidenGorp.
Builders tor industry,
Institutions, hospitals,
manufacturing plants,
government and
developers ol high
technology facilities
lor over 35 years
Edward R. Mardon '41
Douglas R. Mardon '82
280 Lincoln Street
Boston, MA 02134
(617) 782-3743
Hennessey, Frank Carbin, Greg Huber, Avi Orn-
stein, Oren Cheyette. John Trussing. Edwin
McMillan, Charles Sullon, and the proposer.
Smith D. Turner (/dt).
N/D 3. Given a door lamp wilh two bulbs, in
which each socket has a chain which when pulled
will change the on/off stale of only the bulb in
that socket. When the lamp is on. it is difficult lo
determine whether one (and if so which one) or
both bulbs are on. The problem is to find the
shortest sequence of pulls that will turn the lamp
completely off sometime during the sequence
(e.g. if chains arc labeled A and B, AAB0AB fits
the requirements, but a shorter sequence may be
found). Can you generalize your solution to a
lamp of three bulbs, four bulbs, etc.? Is your solu
tion unique?
The following solution is from John Spalding:
The solutions can be generated recursively
starting with the solution for the trivial case of
one bulb. Suppose we number the bulbs 1, 2, 3
. . . instead of lettering them. Then the solution
for n bulbs could be given as the following func
tion f(n) giving a sequence of numbers:
f(D = 1f(n) = f(n - 1), n, f(n - 1)
For two bulbs, the solution is thus 1.2,1 or ABA:
for three bulbs, the solution generated would be
ABACABA; for four, ABACABADABACABA, etc.
When the chains are pulled in this sequence, the
bulbs will cycle through what I think is called a
Gray Code, in which all the possible n-digit bi
nary numbers are generated by only changing
one digit at a time. The number of steps required
for n bulbs is thus 2" - 1, which is what we
would expect. Is the solution unique? Well, no.
Multiple working sequences may be generated by
simply permuting the bulbs' labels—e.g., ABA
and BAB. In addition, labels may be permuted al
intermediate levels of the recursion—e.g., ABA
CABA and ABACBAB are working sequences for
three bulbs.
Also solved by Eric Rayboy, R. Bart, Matthew
Fountain, Winslow Hartiord, John Woolslon, Na
omi Markovitz, Mike Hennessey, Frank Carbin,
Oren Cheyette, Pat Kinney, Mike Strieby, Harry
Zaremba, Howard Stem, Yokichi Tamaka, and
Joe Feil.
N/D 4. A regular hexagon can be inscribed in an
equilateral triangle so that its alternate sides coin
cide with the sides of the triangle. What is the ra
tio between the areas of the hexagon and the
triangle?
Walter Cluett has little trouble wilh this one:
Divide the hexagon into six equilateral triangles
and the answer is 6 to 9 or 1 lo 1 1/2.
Also solved by Eric Rayboy, R. Bart, Steve Feld-man, Howard Stern, Ronald Raines. Winslow
Hartford, John Spalding, John Woolston, Naomi
Markovitz. Mike Hennessey, Frank Carbin. Greg
Huber, Avi Ornslein, Pal Kinney, Harry Zar
emba, Frederic Jelen, Ruben Cohen. Stefan Habs-
burg, Peter Silverberg, George Byrd, James
Reswick, Dick Robnett, Raymond Gaillard. Mi
chael Lamoureux. Man' Lindenberg, Smith D
Turner (/dt), and the proposer, Phelps Meaker.
N/D 5. In the country of Moolah, the national
bank issues new dola bills to replace each bill that
wears out or is lost or destroyed, so there is al
ways a constant number of dolas in circulation.
On January 1, the bank issued a new bill with the
picture of Prince Centime replacing that of the
late Queen Peseta. After one year, they found
that 10/27 of the bills in circulation were the new
variety. After two years, 2/3 of the bills; and after
four years all the bills were the new type. What is
the life expectancy of a dola bill?
Harry Zaremba sent us a lucid solution:
Assume N to be constant number of bills in circu
lation. In terms of N and their common denomi
nator, the fractional amounts of new bills in
circulation during the successive four years were
ION/27, 18N/27, 24N/27, and 27N/27. The number
of old-variety bills that were replaced in each of
the four years was ION/27, 8NV27, 6N/27, and 3N/
27, and their respective average years in circulation
were 1/2, 3/2, 5/2, and 7/2 years per dola. Let A
be the weighted average life expectancy' of the old
bills. Then,
AN = ION/271/2 + 8N/27-3/2 + 6N/27-5/2 +
3N/27-7/2, or
A = 1.574 years per dolla.
Also solved by Eric Rayboy, R. Bart, Matthew
Fountain, Winslow Hartford, Frederic Jelen, and
the proposer, Frank Rubin.
Better Late Than Never
Y 1984. Claes Wihlborg and Mats Ohlin have re
sponded.
M/J 1. R. Bart found two alternative solutions.
JUL 3. Smith D. Turner (/dt) found a simpler so
lution strategy.
A/S 1. Samuel Levitin and Benjamin Rouben have
responded.
A/S 2. Samuel Levitin has responded.
OCT 1. R. Bart, W. Smith, and Richard Hess
have responded.
OCT 2. R. Bart has responded.
OCT 3. R. Bart, Samuel Levitin, Richard Hess,
and R. Morgan have responded.
OCT 4. R. Bart, Richard Hess, Samuel Levitin,
Phelps Meaker, and Altamash Kamal have re
sponded.
OCT 5. R. Bart has responded.
N/D SD 1. Michael Strieby and Dick Robnetl
found alternate solutions.
N/D SD 2. The correct answer is 8, as noted by
Eric Rayboy, David DeLeeuw, R. Bart, Winslow
Hartford, John Spalding, John Woolston, Naomi
Markovitz, Pat Kinney, Mike Hennessey, Mike
Strieby, Frederic Jelen, Ruben Cohen, Stefan
Habsburg, Peter Silverberg, George Byrd, James
Reswick, Dick Robnett, Mike Ober, Walter Cluett.
Roger Allen, Harry "Hap" Hazard, Peter DeFoe.
Victor Christiansen, W. Katz, and Phelps Meaker.
Proposers' Solutions lo Speed Problems
SD 1. East has 12 cards in the majors, most likely
6-6-1-0. She needs the VA for her opening bid. In
order to get back to your hand to draw the last
trump, you must not play a spade! (1) It may be
trumped (true!), and (2) You need the *A to exe
cute a so-called "simple" squeeze play. (What
squeeze is simple al the table?). Ruff a heart and
play all your trumps but one, reaching this four-
card position:
North:
* K85
South:
* A93
A3
East:
*QJ10
V A
After you play your last club, discarding a spade
from dummy. East is rendered helpless.
SD 2. The egg. The first egg from which a
chicken hatched had to have been laid by the im
mediate evolutionary ancestor of the chicken. On
the other hand, the first chicken laid something
like what we have for breakfast—what else, a dinosaur egg? Please forward all correspondence to
the Nobel Institute in Stockholm, Sweden. 1'U bewaiting for it there.
A32 APRIL 1985
PUZZLE CORNER
ALLAN J. GOTTLIEB
How to Computerize Your Engagements
As promised in the October issue, here
is the current backlog of submitted prob
lems in various classes. The shortest de
lay until publication is for speed
problems, where I have a half-year sup
ply. The backlogs for chess, bridge, and
regular problems are each approxi
mately one year. Some confusion has
developed concerning the computer-re
lated problems. These problems will be
clearly identified when presented, and
I did not intend to suggest a preference
for computer calculations over mathe
matical analyses for regular problems.
To date (November 9), I have received
two problems designed as computer-re
lated, one of which appears as F/M 1
below. Thus, the backlog for this class
is just one problem.
Finally, I must apologize to Merle
Smith for misspelling his name in the
October column.
Problems
F/M 1. Alfred Anderson inaugurates the
computer-oriented problems with this
offering; he writes:
Recently I used brute force to solve a
rather interesting computer-oriented
problem. Perhaps one of your readers
would find a more elegant solution.
Management meetings are scheduled on
the second Thursday of each month, ad
ministrative conferences are the third
Friday, and work units have a seminar
on the first Monday. Derive an algo
rithm which will generate a date given
the year, month, day-of-week, and or
dinal week within the month. For ex
ample, if a meeting were scheduled for
the third Friday of August 1984, the al
gorithm would return "August 17."
Note that a meeting on the fifth Tuesday
in March would be fine for 1983, 1985,
and 1986 but not for 1984 (there are only
SEND PROBLEMS. SOLUTIONS,
AND COMMENTS TO ALLAN /.
GOTTLIEB. '67. ASSOCIATE
RESEARCH PROFESSOR AT
THE COURANT INSTITUTE OF
MATHEMATICAL SCIENCES.
NEW YORK UNIVERSITY, 251
MERCER ST.. NEW YORK.
N.Y.. 10012.
four Tuesdays in March of 1984). In this
case the algorithm could return January
0.
F/M 2. William Stein likes to deal with
loosely coupled coins:
Two coins, loosely coupled, are
flipped simultaneously such that ifeither one is heads, the other has prob
ability 7/8 of also being heads, but if any
one is tails, the other is equally likely to
be either heads or tails. Find the prob
ability of each individual coin turning up
heads, and the probability of their both
being heads simultaneously (or prove
that the problem statement and data are
inconsistent).
F/M 3. A geometry problem from Phelps
Meaker:
A horizontal line of length 2a forms
the common base for two isosceles tri
angles. On the near side the triangle is
45° - 45° - 90°, and on the opposite side
75° -75° -30°. Determine the radius of
the circle tangent to all sides of the com
posite lanceolate figure, and locate its
center.
F/M 4. Smith D. Turner J(dt) wants you
to find a four-digit number whose
square is an eight-digit number whose
middle four digits are zero.
F/M 5. David Dreyfuss was attracted to
the following problem:
Consider two dipoles with dimen
sions as indicated. The lower dipole is
fixed, and the upper dipole is con
strained to move along a horizontal line.
(This is roughly the geometry encoun
tered in magnetic stirring.) Find the con
ditions on h and b for which the upper
dipole tends to center (the force is in the
opposite direction of d, the displace
ment from the center line). When does
the motion of the upper dipole approx
imate simple harmonic motion?
Speed Department
SD 1. David Evans has placed a turtle
in each of the four corners of a square
room measuring 3 meters on a side. All
four start moving.at the same instant at
a constant speed of 1 cm./sec, and each
crawls directly toward the turtle to the
left. How long does it take for them to
meet at the center of the room?
SD 2. Steven Bernstein knows a teacher
who brings apples for her students:
Ms. Lang, the third-grade teacher,
wanted to do something nice for the n
students in her class. One day she
brought in an apple which would be
given to a lucky youngster. Her problem
was how to choose the lucky one fairly.
Here is what she decided to do: She said:
"I'll pick a number from 1 to n. The first
one of you to guess the number is the
lucky winner. Let's hear your guesses
in alphabetical order. Aaron, you're
first. Be sure to speak up so everyone
can hear you. Zelda, you'll be last."
"Unfair!", said Aaron. "I'll have all
those numbers to choose from. My
chances of guessing are pretty small.
Wanda, Yolanda, and Zelda will have a
chance to hear everyone else's guesses
so they will have a better chance of win
ning!" "Wait a minute," said Zelda. "1
probably won't even get a chance to
guess because all the other kids go first
and one of them will win before it's even
my turn!" Ms. Lang replied, "Children,
I've thought about it and this procedure
is fair. You all have the same chance of
winning." Is Ms. Lang correct?
Solutions
OCT 1. What is the minimum number of high-card points needed to make a contract of 7 spades
A20 FEBRUARY'MARCH 1985
M.I.T.
A-C
D.E
F.G
H
All items available a, M.U. Student Center. Tech Coop open Mon.-Sa,"American Express welcome. Call toll free: 1-800-79
1-8OO-343-5570 outside Mass.
A. CLEAR ACRYLIC PAPERWEIGHT. PewterM.I.T. beaver mascot mounted inside. $15
B. SOUD BRASS PAPERWEIGHT with M.I.T.seal engraved. $15
C. INSIGNIA KEYCHAIN of brushedpewter. $6
D. SHOT GLASS with MIT insigniaseal. $3.50
E. GLASS BEER MUG with platinum MIT.insignia seal. $6
F. Mil CERAMIC IRISH COFFEE MUG.Gold trim, handle and MIT seal. $I5.5O
G. M.I.T COFFEE CUP. Ceramic with tricolor insignia seal. $5.95
H. CERAMIC PIGGY BANK. Creamy whitegold trim ears and hooves. Scarlet MITseal. $12.95
J. BRUSHED ALUMINUM MUG withpewter insignia seal and clear glassbottom $17
M.I.I ACTIVEWEAR
BY CHAMPION
K. Mil CREWNECK STYLE SWEATSHIRT.Cotton/acrylic fleece in scarlet/whiteseal or grey /scarlet seal. Adult sizes S ML XL. $16 (Not Shown) Hooded Sweat-shirt with pouch pocket, sizes S M LXL. $20
L M.I.T. T-SHIRT with contrasting bandingand insignia seal in grey or maroon adultsizes S. M. L. XL. $8. Youth sizes XS XM. L. $7. ' '
M. M.I.T. BEAVER T-SHIRT in white withscarlet trim. Cotton /polyester blend
adult sizes A. M. L. XL. $8 Youth sizes XSS. M. L. $7.
HARVARD
COOPERATIVESOCIETY
7Z£"*' Vlsa
against the best defense?
Bob Sackheim sent us the following:
I suspect that the minimum number of high-
card points needed to make a 7-spade contract is
five. One possible deal would be as follows (point
cards are shown; all other cards are indicated by
x's, their values being immaterial):
6 x x x x x
A K
¥ AQxx
♦ K
* AKxxxxx
♦ xxxxxxxx
*-
A A ) x x x x
Vxxxxxxx
♦ Q¥ KJ
♦ AQJx
+ QJxx xx"
If West leads the AK, South takes with the A A,
cross-ruffs four hearts and four diamonds, setting
up three hearts in his hand, plus his last spade
for 13 tricks. If West leads a heart. North ruffs,
leads a spade to the A A, then three more hearts
and three more diamonds are cross-ruffed, setting
up three good hearts in South's hand, plus his
two remaining spades for 13 tricks. If West leads
a diamond. South ruffs, plays the AA, then
cross-ruffs four hearts and three diamonds, end
ing in dummy; his last spade, and either three
hearts or three diamonds, are good. If West leads
a club. South ruffs, plays his AA, cross-ruffs four
hearts and four diamonds, ending in his hand;
his last three hearts are good.
Also solved by Tom Harriman, Lester Steffens,
Richard Boulay, Philip Dangel, Alan Robock,
Warren Himmelberger, Winslow Hartford, and
the proposer, Howard Sard.
OCT 2. Given a rotating hollow semi-sphere with
a hole in the bottom, into which a marble is
dropped, find the angular velocity required to
hold the marble exactly halfway between top and
bottom of the sphere. Assume the frictional coef
ficient between marble and sphere is |i and the
radius of the sphere is R.
R m
rsinS
In figure shown,
r = radius of the marble
R = radius of the hemisphere
W = weight of the marble
w = angular velocity of the hemisphere
g - acceleration of gravity
and A is the point of contact between the marble
and the spherical surface at a height R/2 above
the bottom. For equilibrium in a vertical uni
formly rotating plane, the moment of W about
point A will equal the moment of the centrifugal
inertia force W/gy/hc about the same point. Thus,Wr sin 8 = W/g-wVr cos 6, or
i 6. Since
9 = cos-1(r/2)/R = 60°,
then tan B = VJ. Alsax = (R - r)sin 6 = V3/2-(R - r).6ubstitutinfi x and tan 8 into equation (1) yields
w = V2g/(R - r) rads./sec., the required angular
velocity. If the intent of the problem was to have
the center of the marble halfway above the bot
tom, we would have
cos 8 = R/[2(R - r)] and
sin 8 = x/(R — r). Hence, for this condition, the
angular velocity from (1) would be:
w = V2g/R rads./sec.
Also solved by Matthew Fountain.
OCT 3. In the Illinois Lottery Lotto game,-the
player chooses six different integers from 1 to 40.If the six match, in any order, the six different in
tegers drawn by the lottery, the player wins the
grand prize jackpot which starts at SI million and
grows weekly until won. Multiple winners split
the pot equally. For each SI bet, the player must
pick two, presumably different, sets of six inte
gers. Considering the grand prize alone, under
what conditions would it pay, on the average, to
play this game? In the game week ending June
18, 1983, 78 people matched all six winning inte
gers and split the jackpot. Estimate the odds of
this outcome, given that 2 million people bought
SI tickets that week.
There are f ^ I or 3,838,380 possible combinations
which might be chosen. It will prove highly con
venient to round this off to 4 x 10s. The expecta
tion for a lottery participant would be easy to
compute, if we didn't have to worry about multi
ple winners. Choosing 2 out of 4,000,000 combi
nations, someone would have a l/2,000,000th
chance of winning. The prize would have to ex
ceed S2 million before it would pay, on the aver
age, to play the game. But, in fact, the prize
would have to be significantly higher than this,
because there is a good chance of multiple win
ners. The probability that a winner will have to
share his prize is only 1/e, as can easily be
shown. There are, according to information given
later in the problem, 2,000,000 people playing the
game. That is, 4,000,000 combinations are se
lected. Since there are 4,000,000 possible combina
tions, the probability that any given selection
does not match the winner (assuming that each se
lection is equally likely) is '
|(4 x 106) - l]/(4 x 10s).
The probability that all the other selections don't
match the winner is(|(4 x 10s) - l)/4 x 10*)« x 10s.
This is very dose to 1/e since, as is familiar from
calculus,
lim ((n - l)/nj" = 1/e.
So there is a probability of 1 - 1/e that there will
be additional winners. We will need to know the
probabilities of specific numbers of additional
winners, and these can be obtained from the
I'oisson formula:
y
1/e).37 (
.37
.18
.06
.01
negligible
i h
This gives the probability of k additional winners,
where u. is the average number of winners. In the
present case |i is 1, since there are 2,000,000 par
ticipants, each with a l/2,000,000th chance of suc
cess. The following table may be compiled:
Probability
No additional winner 37 (1)
One additional winner
Two additional winners
Three additional winners
Four additional winners
Five or more winners gg
What this indicates is that a winner has a .37
chance of getting all the money, a .37 chance of
getting half the money, etc. What the winner
could expect would be, roughly,
.34 + .37/2 + .18/3 + .06/4 + .01/5 = .63
of the prize money. In order for the lottery to be
a good bet, 63/100 times 1/2,000,000 times the
prize money would have to exceed SI. That is,
the prize money would have to exceed S3.17 mil
lion.
To determine the probability of the June 18,
1983, result, we again make use of the Poisson
formula, with k = 78. This gives the probability
that 78 people will have a given winning combi
nation. Since 78! is on the order of 10'", this is anextremely low probability. An event this unlikely
could not, practically speaking, have happened.
But it did happen. Therefore, something has gone
wrong in the preceding calculations.
The problem lies in the assumption that each
combination is equally likely to be picked by lot
tery participants. In fact, people do not choose
Ventures
andAdventures
Stari-Ups
Management
Consultation
Special Projects
Bridging
Tha Technical
and
Tha Organizational
Tho Bureaucratic
and
The Organizational
Specializing In tho
solution of complete,
difficult, new,
Infroquont, or unique
probloms
Giorgio Plccagll,
'67, Ph.O.
100 Oorado Torrace
San Francisco, CA94112
(41S) 333-5084
Gecnge A.
Roman &
Associates Inc
Architecture Planning
Interior Design
Institutional
Commercial
Industrial
Residential
Site Evaluation
Land Use Planning
Master Planning
Programming
Interior Spaco
Planning
Collages
Hospitals
Medical Buildings
Office Buildings
Apartments
Condominiums
Qoorgs A. Roman,
A.LA.6S
One Gateway Center
Newton, MA 02158
(617) 332-5427
Haley &
AldndvInc.
Consulting Gaotechnlcal
EnQlnocrs ond
Geologists
Soil and Rock
Mechanics
Engineering Geology
Engineering Geophysics
Foundation Engineering
Terrain Evaluation
Engineering Seismology
Earthquake Engineering
Geohydrology
238 Main St.
Cambridge, MA 02142
(617) 492-6460
Hart P. Aldrlch, Jr. '47
Martin C. Murphy '51
Edward B. Klnnsr '67
Douglas Q. Glttord '71
Joseph J. Rlxnor '68
John P. Dugan '68
Kenneth L Rocker 73
Mark X. Haley '75
Robin B. Dill 77
Andrew F. McKown 78
Keith E. Johnson '80
TECHNOLOGY REVIEW A21
Ol * 1 1
Crap.
■ Contract research and
development In
radio frequency,
microwave and
millmeter wave
engineering and
related aroas.
RF and Microwave
Systems Design
Industrial Applications
of Microwave Power
Precision
Instrumentation
Analog and Digital
Electronics
Manufacturing
facilities available
165 New Boston Street
Woburn, MA 01601
(617) 9354460
tier
Our new Macrometer*™
geodetic surveying
system uses GPS
satellite signals to
measure positions and
basollnes with part-
per-mllllon accuracy In
all three coordinates.
Thomas
KDyer,Lnc
DIVISION OF HNTB
Consulting Engineers
Rail Transportation
Thomas K. Dyer '43
1762 Massachusetts Ave.
Lexington, MA 02173
(617) 862-2075
Washington, D.C.
(202) 466-77S5
Chicago, IL
(312)663-1575
Philadelphia, PA
(215) S69-178S
NelsoivCoulson&Associates,lnc
Professional Staffing
Consultants
Contract Engineering
Services
Professional
Recruiting Services
Technical Personnel In
All Disciplines
333 W. rtampdon Ave.
Suite 507
Englewood, CO 80110(303) 761-7660
Other offices In
Albuquerque,
Colorado Springs,
Dallas & Seattle
Paul R. Coulson
PE.-43
President
numbers at random. They choose numbers which
have some significance. Often this is a strictly in
dividual significance (one's age, a date of the
month with special meaning, etc.). But the
choices are sometimes on the basis of a more gen
eral significance. Some numbers (e.g., 7 and, to a
lesser extent, 3) are believed to be lucky. These
and related numbers, such as 21, 33, and 37, will
be heavily over-represented in the selections. By a
reverse psychology, 13 might also be popular.
What were the winning numbers for June 18,
1983? This can be looked up in daily papers of the
region for June 19. When I did look this up, I had
some idea, based on the considerations in the last
paragraph, of what 1 might be seeing. But, even
so, I could hardly believe my eyes. The winning
combination was: 7, 13, 14, 21, 28 and 33. It
would be difficult to imagine a group more ap
pealing to the believer in lucky numbers! Here we
have all the products of 7 between 1 and 40, with
13 thrown in for good measure.
So the answer to the question "What was the
likelihood of the June 18, 1983, result?" is now
apparent, though it doesn't lend itself to exact
quantification. It was quite remarkable (having a
probability of 1/(4 x 10s) that that particular com
bination was the winner. It was not at all remarka
ble that 78 people chose it. The probability of the
complex event (that number winning and being
chosen by 78) is not much less than 1/(4 x 10°).
Let's consider the more general question of
there being some result involving a very large
number (say 50 or more) of winners. I am in
clined to think that this is not extraordinary un
likely. Take the following as a list (obviously not
exhaustive or definitive) of lucky numbers be
tween 1 and 40: 3, 7, 11, 13. 14, 21, 22. 28, 33, 35.
No other winners
One other winner
Two other winners
Three other winners
Probability
.61
.30
.08
.01
37. There are 462 combinations of these
numbers which might be selected. Given that the
total of possibilities is 4 x 10s, it's not likely thai
any of these will again be selected in the foresee
able lifespan of the lottery. But consider the com
binations of five of these numbers with some
other number selected for its significance to an in
dividual. There are (™\ x 29 (about 14,400) ofthese combinations. The chance that one of them
will be a winner is roughly .004 (1 in 250). So one
of these should come up about every five years,
and when it does there may well be a lot of win
ners. Probably not as many as 78, for these com
binations don't have the unique appeal of that
other one, but 50 would not seem too unreasona
ble for some of these combinations.
Ifs now quite clear that my earlier calculation
of what percentage of the prize a lottery player
might expect to collect can only be interpreted as
an average. What a particular player can expect
will vary in relation to the numbers that are se
lected. If he picks two of the supposedly lucky
combinations, his winnings may have to be
shared with 50 others. If the grand prize is at the
$1 million level, each winner's share is slightly
less than $20,000. Recalling that there is still just a
1/2,000,000th chance of winning, this works out
to an expectation of less than \t.
If you deliberately choose nothing but "ordi
nary" numbers, your expectation is somewhat im
proved beyond the level I gave earlier. Let's try a
quite arbitrary assumption, in order to make the
discussion more concrete. Suppose half of the lot
tery entrants make use (consciously) of one or
more of the "lucky" numbers in picking their
combinations. The other half choose their num
bers on the basis of considerations peculiar to
themselves, so that these selections are, in the ag
gregate, random. If you deliberately choose only
"ordinary" numbers, then you are competiting
only with the second group. There is no possibil
ity that you will have to share your prize with
anyone in the first group. Using the Poisson for
mula, we can compute a table of probabilities for
this situation, similar to the table given earlier.
But now |i is just 0.5. We have, on the supposi
tion that you win:
You would stand to receive .61 + .30/2 + .08/3
+ .01/4, or about .79 of the prize. In the case
where the prize is SI million, this works out to an
expectation of 40j (compared to the 1« expecta
tion enjoyed by the other contestant). But note
that under any circumstances, the prize must be
significantly above $2 million before the lottery
becomes a good bet.
Also solved by Frank Carbin, Warren Himmel-
berger, Matthew Fountain, and the proposer. Jon
athan Hardis, who sent us a particularly complete
solution. Mr. Hardis seems to be quite an author
ity on the Illinois lottery. Since he has a Chicago
address, one might conjecture that some of his
knowledge comes from empirical study.
OCT 4. An associate research professor walks
into his office one morning and says to his secre
tary, "I had three dinner guests last night. The
product of their ages was 2450. The sum of their
ages was twice your age. Can you tell me their
three ages?" Ten minutes later his secretary came
to him and said that the problem could not be
solved. He said, "You are right. I will now tell
you that I was the oldest one there." The secre
tary was then able to tell him the ages of.the
three dinner guests. What are the ages of the din
ner guests, her age, and the professor's age?
The following solution is from Tom Gallahan:
The first thing to do is make a list of the possible
combinations of ages. Once the secretary has
done this he/she can eliminate the ones that do
not add up to twice his/her age. If there were
only one set of ages that fit the criteria it would
be the answer. This is not the case because he/she
cannot solve the problem. There must be more
than one choice. The fact that the professor is the
oldest one there must distinguish between the
possible choices so that the secretary can solve
the problem.
In making a list it helps to make a factor tree,
since three ages's product is 2450.
2450
\98
2 49
/\
25
/ \5 5
2450 = 2x5x5x7x7 7 7
Ages of Resulting age
guests of secretary
1 2, 5,245 126
2 2, 7, 175 92
3 2, 25,49 38
4 2, 35, 35 36
5 5,5,98 54
6 5, 7, 70 41
7 5, 10,49 32
8 5. 14,35 27
9 7, 7, 50 32
10 7, 10,35 26
11 7, 14, 25 23
You must also realize that one or two of the
guests may be one year old:
12 1,25,98 62
13 1,35,70 53
14 1,49,50 50
15 1, 1, 2450 1226
None of the guests may be 0 years old because
the product of the guests' ages in that case will
always be 0.
For ease of explanation 1 have numbered the
sets of ages. All of the resulting secretary's ages
are distinct except 7 and 9. One of these must be
the correct answer; thus we now know that the
secretary is 32 years old. In case 7 the professor
must be 50 years old or older to be the oldest one
there. In case 9 he must be 51 or older. If the pro
fessor is 51 or older the secretary can't choose be
tween the sets. The professor must be 50 years
Ml FEBRUARY/MARCH 1985
old and 7 must be Ihe correct set. logically we
could have eliminated 1, 2, and 15 but this is un
necessary.
Also solved by Michael Jung, Matthew Foun
tain, David Griesedieck, Tom Harriman, Naomi
Markovitz, Richard Boulay, Fernando Saldanha,
Jerry Sheldon, Clarence Cantor, Steve Feldman,
Dennis Loring, Ronald Raines, James Michelman,
Avi Omstcin, Tso Yce Fan, John Rosendahl, Ray
mond Caillard, E.R. Foster, Danny Mintz, Roy
Levitch, Miriam Nadel, Tom Lydon, Myles Fried
man. Pierre Heftier, Frank Carbin, VVinslow Hart
ford, Harry Zaremba, and the proposer. Merle
Smith.
OCT 5. Given an irregular polygon of n sides, in
which sequence should Ihe sides be arranged and
how should the comer angles be determined to
give the greatest area?
The following solution is from the proposer,
Irving Hopkins:
Consider that the polygon is the horizontal
cross-section of a vertical open-topped water
slandpide, the rectangular sides of which are
hinged together along their vertical edges. There
is no friction in the hinges, and Ihe lower end of
the vessel rests in a friction-free manner on a hor
izontal plate with no leakage of water. Gravity
causes the water to fall until the sides have
moved to what must be the enclosure of maxi
mum area. When movement of the water has
ceased, the pressure at any depth is uniform in
all directions. The stress in the sides depends on
the depth, but the geometry does not vary with
the pressure. Assume a depth at which a side of
width L and a suitable vertical dimension is sub
ject to a force 2L. At the junction of L, and L2 the
forces are as shown, where t, is the tension in L,
and t: that in L:.
Length L, l|
At the hingeptn, Ihe forces parallel to L, balance
if
t, = tjCOsA, + L2sinA, (1)
and the perpendicular forces if
L, + LjCosA, = t;SinA, (2)
Going on to other comers, we have:
tj = t)COsA2 + L>sinA2 (3)
Lj + LjCOsA2 = t}sinAj (4)
(2n - 1)t. = tgCOsA, + L,sinA,,
L, + L]CosA, = t|SinA,. (n)
From the even-numbered equations above:
t, = (L, + L,cosAJ/sinA.
t2 = (L, -t- LjCosA,).'sinA,
t, = (L,_, + UcosA, - l)/sinA,.
Substituting these values of t in the odd-num
bered equations, we get:
(L,cosA. + L,)/sinA. = (L,cosA, + LJ/sinA, (I)
(L^osA) + Li)/sinAi = (L2cosA2 + Ls)/sinA2
(H)
(UcosA,., + L,.|)/sin A,,., = (L,cosA, + Lt)/
sinA,. (Ill)
We now have n equations from which to find the
values of angles A, to A.. But there is one more
requirement: the sum of Ai to A, must be 360s.
Assume a value for A] and let Q equal the left-
hand side of (IT). Square both sides of (IT), which
becomes
cosJAi(L,! + Q2) + 2LjLjCosA2 + (Lj2 - Q2) =O.
This is a quadratic equation in cosA? from which
A] may be found, and so on. For a pentagon with
sides 3, 5, 7, 9 and 11 in length the angles were
found to be
Goldbeig-Zoino &Associates Inc
Gaolochnlcnl-
Goohydrologlcal
Consultants
Foundation/Soil
EngineeringSite Planning ft
Dovolopmont
Soil/Rock Testing
Geotochnlcal
Instrumantatlon
Testing of Construction
Engineering for
Latsral Support
Systems
Rock Engineering
Groundwater
Englnoerlng
Underground
Construction
D. T. Goldberg, '54
W. S. Zolno, '54
J. D. Guertin, '67
M. J. Borvenlk, 76
M. D. Bucknam, '81
N. A. Campagno, Jr. '67
f. W. Clark, '79
W. E. Hodge, 79
W. E. Jaworekt. 73
C. A. Llndberg. 78
R. M. Simon, 72
E. I. Stolnberg, '60
T. vonRosenvlnge IV, '80
The GEO Building
320 Noodham St.
Newton Upper
Falls, MA 02164
(617) 467-8840
Edward R
MaidenGap.
Bulldors for Industry,
Institutions, hospitals,
manufacturing plants,
government and
developers of high
technology facilities
lor over 35 years
Edward R. Mardon '41
Douglas R. Marden '62
280 Lincoln Street
Boston, MA 02134
(617) 782-3743
JamesGoldstein&Partners
ARCHITECTSENGINEERSpi &MKIPQQ
S. James Goldstein '46
Eliot W. Goldstein '77
225 Mlllburn Avanue
Mlltburn, NJ 07041
(201) 467-8840
SERVING THE SCIENCE COMMUNITY SINCE 1SS1
FACILITIESResearch &
Development
Education & Training
Management &
Support
Manufacturing &
Warehousing
FOR HIGHTECHNOLOGY
FIELDS
Biochemistry
Chemical Englnoorlng
Chemistry
Computer Science
Electronics
EMI Structures
Hazardous Materials
Information Science
Laboratory Animals
Lasers
Medical Sciencas
Microotoctronlcs
Monoclonal
Antibodies
Parllclo Accelerators
Pulp & Paper
Recomblnant DNA
Solid State Physics
Telecommunications
Toxicology
Wind Tunnels
Pi(xagli&Associates
Corporate
Hoolih Care
Consultants
Giorgio Plccngll,
Ph.D., '67
100 Dorado Terrace
San Francisco, CA94112
(415) 333-5084
IheCodman
Gompany,Inc
Industrial and
Commercial Real Estate
Mark Gottesman '70
(M.C.P.)
211 Congress Street
Boston, MA 02110
(617)4234500
PaulEDutelle&Company
Inc
Roofers and Metal
Craftsmen
153 Poor! Street
Newton, MA 02158
TECHNOLOGY REVIEW A23
LordElectric
CompanyLnc
Ekictrical contractors
to tho nation slnco
Hoadquartors:
45 Rockofotler Plan
Now York. N.Y. 10111
Offices In 16 principal
cttlos throughout tho
U.S. and Puorto Rico
Boston Office:
66 Coolldga Ave.
Wstertown, MA 02172
(617) 926-5500
Alexander W. MoHat. Jr.
H.H.
Hawkins&SansCompany
BulldlrtQ contrsctofs
Sloven H. Hawkins, '57
188 Whiting Street
Hlnglum, MA 02043
(617)749-6011
(617) 749-6012
Koalth Care Consultant*
Design, Construction,
Management
Subsidiaries:
Charles N. Oobes &
Assoc. Inc.
Alma Nelson Manor Inc.
Park Stratrtmoor
Corporation
Rockford Convalescent
Center Inc.
Chambra Corporation
Chariot N. Oobes '35
5666 Strathmore Drive
Rockford, IL. 61107
75916367
37 590272"
57.838328
68.359986° X ,2396886.
80 4 25299"
108.22973"
A,
A2
A,
A,
A,
37.590°
57.838°
80.425°
108.230"
75.916°
Could such polygons be circumscribed by circles?
If so, each side of the polygon is a chord of the
circle, whose center must lie at the point of inter
section of the perpendicular bisectors of the
chords. Consider two adjacent sides of the poly
gons, say L, and Lj, with angle A,.
Taking
c = (LV2)sinA, + qcosA, and horizontally
L,/2 = qsinA, - (Lj/2)cosA,.
From these equations we find that
c = [Lj/2 + (L|/2)cosA,]/sinA,,
and the radius R12 determined from sides L, andLjis
R!2 = (c3 + (L,^)2)"2
|(j) (.ajcos
which boils down to
4R»2 = (L,2 + 2L,LjCOsA, + L22)/sin2A, (IV)The existence of a circumscribing circle depends
on R,2 = Ra = . . R.,. By analogy with (IV),
4Ra! = (Lj2 + 2L2LjCosA2 + L32)/sin2A2 (V)Squaring both sides of (11), we get
(L,2 + 2L,L2cosA, + LjWA,)/sin2A, = (L,2 +2LjL3cosA2 + L^cos^j/sin'Aj. (VI)Subtracting the left side of (VI) from the right sideof (IV), and the right side of (VI) from the right
side of (V), we get
LjJ(l - cos2A,)/sin2A, = L,2, and
L22(l - cos2A2)'sin2A2 = W-The right sides of (IV) and (V) are therefore equal;
hence RI2 and Ro and all the other R's are equal.
The area of each triangle is easily found by
Area = (L/2)[R2 - (L/2)2]"3. The sequence inwhich the sides are arranged is immaterial, equiv
alent to carelessly cutting a pie and then shuffling
the pieces.
The pentagon described is shown above.
Each side subtends an angle at the center [B, =
2arcsin(L/2R)), and the sum of these must equal
360°. The easiest way to find the radius giving B
= 360° is cut-and-try. This may fail if a very long
side, U, causes the center of the circumscribing
circle to be outside the enclosure. In this case,
solve for
Also solved by Matthew Fountain, Tom Ham-
man, Winslow Hartford, and Harry Zaremba.
Better Late Than Never
M/J 4. Andre Schmitz found a simpler way to
present the solution.
A/S 4. Dick Swenson has responded.
Proposers' Solutions to Speed Problems
SD 1. Five minutes. Since the turtles move at
right angles to each other, an approached turtle's
motion does not contribute to the distance the approaching turtle must travel.
SD 2. Yes, Ms. Lang is correct. Each child has
probability 1/n of winning. Aaron's chance is 1,'n
because he has n numbers to choose from. The k-
th child will win if each of the k - 1 children that
went before him/her guessed incorrectly and the
k-th guesses correctly. The probability of this happening is:
[(n - l)/n|[(n - 2)/(n - 1)] . . . (n - k + l)/(n- k + 2)[l/(n - k + 1)] = I/n.
Notice that in order for this to work it is neces
sary for each guess to be heard by the children to
follow. In this way the disadvantage of going
near the end is exactly compensated by a narrow
ing down of choices.
A24 FEBRUARY/MARCH 1985
Ct ' 1 1
Goip.
Contract research and
development In
radio frequency,
mlcrowovo and
mllimeter wavo
engineering and
rotatod areas.
RF and Microwave
Systems Design
Industrial Applications
of Microwave Power
Precision
Instrumentation
Analog and Digital
Eloetranlcs
MsnufsctunnQ
facilities available
185 Now Boston Street
Wobum. MA 01801
(617) 935-9460
ier
Our new Macromotor1™
geodotlc surveying
system uses GPS
satellite signals to
measure positions and
baselines with part-
per-mllllon accuracy In
all three coordinates.
PUZZLE CORNER
ALLAN J. GOTTLIEB
100 Ways to Say 1985
BoyleErigineerijGap.
Engineers/Architects
Complete Professional
Services:
Water Supply
Pollution Control
Architecture and
Landscape
Architecture
Highways and Bridges
Dams and Reservoirs
Electrical-Mechanical
Engineering
Environmental Science
Computer Sciences
Agricultural Services
Management ond
Administration
Haley &Aldnch,
Inc
Consulting Geotechnlcal
Engineers and
Goologlsts
Soil and Rock
Mechanics
Engineering Geology
Engineering Geophysics
Foundation Engineering
Terrain Evaluation
Engineering Seismology
Earthquake Engineering
Geohydralogy
238 Main St.Cambridge. MA 02142
(617) 492-6460
Thomas S. Maddock '51
1501 Quail Street
P.O. Boi 7350
Newport Beach, CA
92660
(714) 752-1330
Harl P. Aldrlch, Jr. '47
Martin C. Murphy '51
Edward B. Klnrtar '67
Douglas G. Gifford '71
Joseph J. Rlxner '68
John P. Dugan '68
Kenneth L. Reckor '73
Mark X. Haley °75
Robin B. Dill '77
Andrew F. McKown '78
Keith E. Johnson 'B0
This being the first issue of another
year, we again offer a "yearly
problem" in which you are to ex
press small integers in terms of the digits
of the new year (1, 9, 8, and 5) and the
arithmetic operators. The problem is for
mally stated in the "Problems" section,
and the solution to the 1984 yearly problem is in the "Solutions" section.
Problems
Y1985. Form as many as possible of the
integers from 1 to 100 using the digits
1, 9, 8, and 5 exactly once each and the
operators +, -, x (multiplication),/(di-
vision), and exponentiation. We desire
solutions containing the minimum num
ber of operators; and, among solutions
having the minimum number of oper
ators, those using the digits in the order
1, 9, 8, and 5 are preferred. Parentheses
may be used for grouping; they do not
count as operators.
JAN 1. Our next problem is the last
member of Emmet Duffy's collection of
seven-card bridge problems. For the cur
rent challenge, South is on lead and is
to take six tricks against best defense
with hearts as trump:
A
V
♦
J8
J10 9 8
5
*
♦
♦
A
V
♦
*
983
10
AQ
AK—
—
AQ
2
7
43
A
V
♦
10
6
K6
KJ6
JAN 2. Bruce Calder, after working on
1983 N/D 4, sent us the following spin-
off, a problem demonstrating the ele
gant subtlety of Newtonian mechanics:
A smooth, rigid, and circular wire
SEND PROBLEMS, SOLUTIONS.
AND COMMENTS TO ALLAN /
GOTTLIEB. 67. ASSOCIATE
RESEARCH PROFESSOR AT
THE COURANT INSTITUTE OF
MATHEMATICAL SCIENCES.
NEW YORK UNIVERSITY, 251
MERCER ST.. NEW YORK.
NY.. 10012.
hoop hangs from a rigid support by an
ideal, extensionless string. Two small
beads slide along the hoop (like beads
of a necklace) with negligible drag and
friction. The beads are slid to the top of
the hoop and released. How massive
must each bead be to spontaneously lift
the hoop?
JAN 3. Here's one John Rule dug out of
the file where he keeps interesting prob
lems encountered from various sources:
A man received a check calling for a
certain amount of money in dollars and
cents. When he went to cash the check,
the teller made a mistake and paid him
the amount which was written in cents
in dollars, and vice-versa. Later, after
$3.50, the man suddenly re
alized that he had twice the amount of
money the checked called for. What was
the amount on the check?
JAN 4. Our last regular problem is from
Floyd Kosch:
yyy
MA JANUARY I
Frank E. Reeves, '24; August 26. 1984; 1661 Mo
hawk St., Los Angeles, Calif.
Lloyd R. Rogers, '24; December 31, 1984; 2129 Pol
Springs Rd., Lulherville, Md.
Gavin Watson, '24; September 21, 1984; PO Box
1099, Arizona Bank, c/oTrust Dept., Sun City, Ariz.
Francis E. Field, '25; July 6, 1984; 32 Buena Vista
Rd., Biltmore Forest, Asheville, N.C.
Harry Newman, '25; July 1984. 220 W Jersey St.,
Elizabeth, NJ.
Robert A. Nisbct, '26; July 21, 1984; 25 Barrows
Terr., Stratford, Conn.
Robert N.C. Hesscl, '27; July 20, 1934; 22 Saxib Rd.,
Worcester, Mass.
Cordon E. Thomas, '27; September 1984; 20 Terrene
Ave., Nan'ck, Mass.
Victor J. Decode, '28; July 9, 1984; Ocean Club No.
1109, 4020 Gait Ocean Dr., Fort Lauderdale, Fla.
Prescott D. Crout, '29; September 25, 1984; 9 Pine-
wood St., Lexington, Mass.
Richard V. Does, '29; 1984; 27 Hamilton Ave., Ded-
ham. Mass.
Emmette F. fcard, '29; September 3, 1984; RT 1 Box
159, Hazlehurst, Miss.
Delmer S. Fahrney, '30; September 12, 1984; 10245
Vivera Dr., La Mesa, Calif.
Charles G. Habley, '30; September 8,1984; 210 Mid-
vale St., c/o Felker, Falls Church, Va.
Ellas Klein, '30; January 17,1984; PO Box 115, Half
way House, Transvaal, S Africa.
Harry W. Poole, '30; September 7, 1984; 1201 Mot-
tron Dr., McLean, Va.
Reuben Roseman, '30; March 1984; 520 Wyngate
Rd., Timonium, Md.
Glenn Goodhand, '31; September 11. 1984; 6307
Stoneham Ln., McLean, Va.
Ernest H. Lyons, Jr., '31; 1984; 23871 Willows Dr.
No. 350, Laguna Hills, Calif.
Heinrich W. Weitz, '31; August 28, 1981.
Lucien B. Curtis, '32; 1984; 24 Park St., Brandon,
Vt.
Edwin Allen Newcomb, '32; September 14, 1984;
Stillwater Health Care, Stillwater Ave., Bangor,
Maine.
Gerner A. Olsen, '32; April 26, 1984; 5 B 7 Glen
Ave., Scotia, N.Y.
Frank Der Yuen, '33; July 23, 1984; 1565 Kalaniiki
St., Honolulu, Hawaii.
Gordon A. Danforth, '34; September 19, 1984; PO
Box 444, Richmond, 111.
George E. Agnew, '35; January 3, 1984; 17881 Corte
Emparrado, San Diego, Calif.
Robert Kenneth Bullington, '37; May 30. 1984; 36
Glenwood Rd., Colts Neck, N.J.
Robert D. Morton, '37; October 14, 1984; 82 Sunset
Farm Rd., West Hartford, Conn.
Carl I. Shulman, '38; October 15, 1984; 41 Bethune,
New York, N.Y.
Philip W. Constance, '39; March 31, 1984.
David B. Parker, '40; March 30. 1980.
Bjom Lund, '41; September 12, 1984; 194 Niantic
River Rd., Waterford, Conn.
Neil D. Cogan, '42; August 29, 1984; 47 So Ridge-
land Rd., Wallingford. Conn.
Andrew L. Johnson, '43; October 13, 1984; 211 En-
glewood Ave., Newcastle, Penn.
Donald T. Cloke, '45; July 21,1984; PO Box 173. La
Grange, Maine.
Arthur C. Nisula, '46; September 3, 1984; 907 Ma
rina Dr. No. 401, North Palm Beach, Fla.
Elmer B. Sampson, '47; September 7, 1984; RR 3Box 1368, Leesburg, Fla.
Stanley H. Southwick, '51; April 20. 1984; 935 Ma
ple St., Friend, Neb.
James W. Burch, '53; April 10, 1984; 6 Taney Ave.,
Annapolis, Md.
John M. Houston, '55; August 31, 1984; 1302 RoweRd., Schenectady, N.Y.
Robert D. Alter, '56; March 21, 1984; 258 Woodland
Rd., Highland Park, 111.
Harvey J. Baker, '72; September 9,1984; 3122 North
ampton St. NW, Washington. D.C.
Otto K. Soulavy Burchard, '83; September 1984;
Calle Caurimare, Res Parque, Colinas De Bello
Monte No. 62A, Caracas, Venezuela.
R. Nicolas Harrison, '84; October 6, 1984; 24 Elm
Grove Rd., Ealing London. England.
Edward R.Mantel
Gup.
Builders lor Industry,
Institutions, hospitals,
manufacturing plants,
government anddovolopsrs of high
technology facllltias
lor over 35 years
Edward R. Marden '41
Konnoth fl. Hoffman 78
Oouglas R. Marden '82
280 Lincoln Stroet
Boston. MA 02134
(617) 782-3743
Geoigp A.
Raman &Associates Inc.
Architecture Planning George A. Roman,
Interior Design A.I.A. '65
Institutional One Qateway Center
Commercial Newton, MA 02158
Industrial (617)332-5427
Residential
SKe Evaluation
Land Use Planning
Master Planning
Progrsiliming
Interior Space
Planning
Collegos
Hospitals
Medical Buildings
Office Buildings
Apartment*
Condominiums
Syska&
HennessyInc
Engineers 5901 Green Valley
Circle
Mochsnlcal/Etoctrlcal/ Cu(yer CitySanitary Los Angeles, CA
80230
John F. Hennessy '51
11 Wool 42nd St.
Now York, N.Y.
10038
840 Memorial Dr.
Cambridge, MA
02139
575 Mission St.
San Francisco, CA
94105
Goldbeig-
Zaino &Associates Inc.
Geotechnlcal- D. T. Goldberg, '54
Goohydraloglcal W. S. Zolno, 'S4
Consultant! J. D. Guortln, '67
Foundation/Sell M. J. Barvenik, '76*
Engineering M. D. Bucknam, '81
Site Planning & N. A. Campagna, Jr. °67
Development F. W. Clark, '79
Soil/Rock Testing W. E. Hadge, '79
Geotechnlcal W. E. Joworakl, '73
Instrumentation C. A. Undberg, *78
Testing of Construction R. M. Simon, '72
Engineering for E. I. Steinberg, -so
Lateral Support T. vonRosenvlnge IV. '80
Systems
Rock Engineering The GEO Building
Groundwster 320 Noodham St.
Engineering Newton Upper
Underground Falls, MA 02164
Construction (617)467-8840
DebesCaip.
Hoalth Care Consultants
Doslgn, Construction,
Management
Subsidiaries:
Charles N. Debes &
Assoc. Inc.
Alms Nelson Manor Inc.
Park Strathmoor
Corporation
Rockford Convalescent
Center Inc.
Chambro Corporation
Charles N. Debes '35
S888 Strathmore Drlvo
Rockford, IL. 61107
EEHawkins&Sons
Company
Building contractor*
Steven H. HflwkJns< '57
186 Whiting Street
Hlrtgham. MA 02043
(617)749-6011
(617)749-6012
-
TECHNOLOGY REVIEW AH
A rigid arm pivots around the fixed
point A. At the end of the arm is a fol
lower (B) which runs in a curved track.
The track pivots about the fixed point
C. If AB = AC = r, find the shape of
the track such that its slope at C is al
ways vertical.
Speed Department
SD 1. Here's one from Smith D. Turner
(/dt):
Bill and Joe are to be paid $10 to wrap
and address a pile of packages. Joe ad
dresses one while Bill wraps one, but
Bill addresses three while Joe wraps one.
How should the $10 be divided between
them?
SD 2. We end with a bridge quickie from
Doug Van Patter:
South:
4k AK74
V KJ3
North:
AJ6
V AQ5
♦ A2 ♦ KJ54
+ AQ9 862 ♦ K 5
You are declarer (South) in a six-no-
trump contract. The opening lead of the
♦ 10 is taken with your ^J. Now you
wish you were in a grand slam. When
you cash the *K, West shows out. Can
you find any reasonable chance of still
making your contract?
Solutions
Y 1984. This is the same problem as Y 1985 (see
above) with only the one digit changed.
The following solution is from Harry Zaremba;
he points out that 36 numbers, shown with aster
isks, use the digits in the same order as the year
1984:
37 = (1 + 4) x 9 - 8
"38 = 19 x 8/4
39 = 48 - 9 x 1
40 = 41 - 9 + 8
•41 = 1 x 9 + 8 x 4
•42 = 1 + 9 + 8 X 4
43 = 91 - 48
•1 = 1*2 = 4/(18.-9)
•3-= 1* + 8/4•4 = 1-9 + 8 + 4
•5=1x9-8 + 4
•6=1+9-8 + 4
•7 = 19 - 8 - 4
•8=1 + 9-8/4
9 = 9xlH
10 = 9 - 1 + 8/4
•11 = 1 x 9 + 8/4
•12 = 1 + 9 + 8/4
13 = 94 - 81
•14 = 1 + 9 + 8 - 4
•15 = 19 - 8 + 4
16 = (9 - 1) x 8/4
•17 =19-8/4
•18 = 1 x 9 X 8/4
•19 = 1 + 9 x 8/4
•20 = (1 + 9) x 8/4
•21 = 19 + 8'4
•22 = 1+9 + 8 + 4
•23 = 19 + 8 - 4
24 = 41 - 9 - 8
25 =
26 =
27 =9x4-8-1
28 = 9x4-8x1
29 = 48 - 19
30 =
•31 = 19 + 8 + 4
32 = 81 - 49
•33 = 1° + 8 x 4
34 =
35 = 9 X (4 - I) + f
36 = 81/9 x 4
•44 = (19 - 8) x 4
45 = 81 - 9 x 4
46 =
47 = 48 - 1*48 = 89 - 41
49 - (1 + 4) x 8 -r 9
50 = 49 + I8•51 = 19 + 8 x 4
52 =
53 = (l+4)x9 + 8
54 = 18 + 9 x 4
55 =
56 = 48 - 1 + 9
57 = 98 - 41
58 = 41 + 9 + 8
59 = 91 - 8 x 4
60 = (9-l)x8-4
61 =
62 =
63 = 18 x 4 - 9
64 = (9 + 8-l)x4
65 = 84 - 19
66 =
67 = 48 + 19
•68 =1x9x8-4
•69 = 1 + 9 x 8 - 4
70 =
71 = 9 x 8 - 1*•72 = (1 + 9 + 8) x 4
73 = 9 x 8 + I4
74 = 84 - 9 - 1
75 = 89 - 14
•76 = 1 - 9 + 84
•77 =1+9x8 + 4
78 =
79 = 91 - 8 - 4
80 = (I4 + 9) x 8*81 = 1 x 91*11
•82 = 1 + 9<M>
83 = 84 - 1*
84 = 98 - 14
•85 = 1' + 8486 = 81 + 9 - 4
87 = 91 - 8 + 4
88 = 89 - I4
89 = 91 - 8/4
90 = 18 x (9 - 4)
91 = (9 + 4) x (8 - I)
92 = 89 + 4 - 1
•93 = 1 x 9 + 84
•94 = I + 9 + 84
•95 = 1 + 98 - 4
96 ■= (9 + 4 - 1) x 897 = 98 - I498 = 98 x I4
99 = (8 + 4 - 1) x 9100 =
Also solved by Avi Ornslein, George Aronson.
Harry (Hap) Hazard, Phelps Meaker, Joe Feil, Peter
Silverberg, Allan Tracht, and A. Holt.
A/S 1. Given the situation shown. While to move
and win.
Bert Daniels had a little trouble with this one:
1. N-B8, K-Rl (1 P-N3 loses to 2. P-B6, N-B2,
3. Q-K8 mate; while 1 K-Bl loses to 2. Q-K7
ch and Q-K8 mate);
2. Q-Q8 ch, N-Nl;
3. N-Q6 and the threat of smothered mate wins the
queen.
Also solved by Matthew Fountain, Kenneth Bern
stein, R. Hess, Avi Omstein, David Evans, David
Detlefs, George Aronson, Ronald Raines, Philip
Dangel, and the proposer, Robert Kimble.
A/S 2. An ordinary combination padlock requires
three ordered numbers to open, each between 0 and
39, inclusive. Thus there are 64,000 possible com
binations. If it is known that the sum of the three
numbers is 58 and the sum of the individual digits
of all three numbers is 13, how many combinations
are possible? If each of these possible combinations
is equally likely, what is the probability that the
second number is 34?
Many readers submitted computer programs that
calculated all possibilities. I preferred analyses that
reduced the number of possibilities to a manageable
level. Matthew Fountain actually submitted both a
program and an analysis. Here is the latter:
Let A equal the sum of the tens digits of the three
ordered numbers and B equal the sum of the units
3,2,0
3,1,1
2,2,1
Total
Permutations
6 (2)
3 (1)
_3 _
12 (3)
( ) indicates number of
permutations with 3 in
second position.
8,0.0
7.1,0
6,2,0
6,1,1
5,3,0
5,2,1
4,4,0
4.3,1
4.2.2
3,3,2
Permutations
3
6
6
3
6
6
3
6
3
3
(2)
(2)
(1)
Total 45 (5)
( ) indicates number of
permutations with 4 in
second position.
Nelson,Goulson&
Associates,Inc
Professional Staffing
Consultants
Contract Engineering
Servlcaa
Professional
Recruiting Services
Technical Personnel In
All Disciplines
313 W. Hompden Ave.
Sulto 507
Englewood, CO 60110
(303) 761-7680
Other offices In
Albuquerque,
Colorado Spring),
Dallas & Seattle
Paul R. Coulson
PE/43
President
PaulEDutelle&
GompanyInc
Roofers snd Metal
Craftsmen
153 Pearl Street
Newton, MA 02158
LordElectricCompany
Inc.
Electrical contractors
to the nation since
Headquarters:
45 Rockefeller Plaza
Now York, N.V. 10111
Offices In 16 principal
cities throughout the
U.S. and Puerto Rico
Boston Office:
86 Coolidge Ave.
Watertown, MA 02172
(617) 926-5500
Alexander W. Mortal, Jr.
TECHNOLOGY REVIEW A25
AlbanyInternationalResearchCo.
Contract RAD
for Industry in
Polymors
Fibers
Textiles
Plastics
Composites
Interpretive Testing
Chemical Analysis
Consulting
Expert Testimony
0. R. Petterson '59
E. R. Kaswoll '39
M. M. Platt '42
R. B. Davis '66
0. S. Brookateln '76M. A. Konney '83
1000 Providence
Highway
Dedham, MA 02026
617-326-5500
J. R. Dent '79
R. E. Erlandson '48
0. M. Boars 'B0
F. A. DITaranlo '83
LEA Gioup
UNENTHALEISENBERQ
ANDERSON, INC.ENGINEERS
Building Design
Environmental
Engineering
Sne/CMI DesignRoofing Technology
Consultants to
Industry, commerce,government and
Institutions.
75 Kneoland Street
Boston, MA 02111
(617)4264300
New York, NY
(212) 509-1922
Eugene R. Elscnbera'43
Louis Rexroat
Anderson '50
William S. Hartley '52
David A. Peters 77MSCE
Venfinesand
Adventures
Start-Up*
Management
Consultation
Special Protects
Bridging
The Technicaland
The Organizational
The Bureaucraticand
The Organizational
Specializing In the
solution ol complete,
difficult, new,
Infrequent, or uniqueproblems
Giorgio Plccagli,■67. Ph.D.
100 Dorado Terrace
San Francisco, CA94112
(415)333-5084
Number
of sides
4
6
8
12
20
Edges
each
side
3
4
3
5
3
Total
edges
12
24
24
60
60
Number
of
dihedralsg
12
12
30
30
Dihedralsmeeting
at each
apex
4
4
3
5
NumbernfUl
apices
AV
20y
digits of the three ordered numbers. Then A t B= 13 and 10A + B » 58, with soluHons A = 5 andB =■ 8. The table at the bottom of page A25 showsthat A can be decomposed into three ordered digits,none exceeding 3, in twelve ways, and B can be
decomposed into three ordered digits 45 ways. Thetotal combinations are 12 x 45 = 540. Those with34 as second number are 3 x 5 = 15. The probabilitythat 34 is the second number is 15/540 = 1/36.Also solved by Kenneth Bernstein, Richard Hess
Avi Omstein, David Evans, David Detlefs, GeorgeAronson, Dennis Sandow, Rita Carp, Gerry Cross-man, Harry Zaremba, Richard Marks, WinslowHartford, Yale Zussman, P. Jung, Frank Carbin,Steve Feldman, Aaron Hirschberg, Dave Mohr,Alfred Anderson, Thomas Stowe, and the proposerJohn Prussing.
A/S 3. Fill in the missing entry in the table abovepertaining to regular polyhedra. Can the values inthe last column be determined by a formula?
The table contained two typos: a square has eightapices and three dihedrals meeting at each apex.These errors did not seem to cause much trouble.In particular, Avi Omstein submitted the following:The missing number is 12 apices for the icosa-
hedron. The number of apices is given by the following:
(number of sides)(edges on each side)/(dihedralsmeeting at each apex)or
(total edges)/(dihedrals meeting a I each apex)or
2(number of dihcdrals)/(dihedrals meeting at eachapex).
Also solved by Matthew Fountain, Kenneth Bem-slein, Richard Hess, David Evans, David DetlefsGerry Grossman, Harry (Hap) Hazard, Harry Zaremba, Richard Marks, Winslow Hartford YaleZussman, and Albert Mullin.
A/S 4. Find infinitely many positive integers n notcontaining the digit zero such that nJ - 1 containsjust two digits neither of which is zero The dicitsmay be repeated.
Jerry Marks sent us three patterns:7
67
667
5667
66667
5
65
665
6665
48
4488
444888
44448888
4444488888
24
4224
442224
44422224
34
334
3334
33334
15
1155
111555
11113555
1111155555
66665 444422224
,-> RiC'j™tHcSS "a,S a Proof of one of these patterns:(2 x 10-/3 + 1/3)' - 1 = 4 x 10*V9 + 4 x 10"/9+1/9-1
ThisSpaceAvailable
For your
advertising
message
Coll:
Poter Gellatly
Technology Review(617)253-8290
= 4(102n - l)/9 + 4(10" - l)/9 + 4/9 + 4/9 - 8/9= 2n 4's + n 4's
= n 4's followed by n 8's.
Also solved by Kenneth Bernstein, David Evans,David Detlefs, Dennis Sandow, Gerry Grossman,Harry Zaremba, and the proposer, Matthew Fountain.
A/S 5. Given a triangle ABC, draw its incircle andconsider triangle DEF determined by the points oftangency. Show that the area of triangle DEF is (r/
d)A, where r is the radius of the incircle, d is thediameter of the dreumcirele, and A is the area oftriangle ABC.
We give two different solutions, the first fromKenneth Bernstein and the second from PhelpsMeaker:
Bernstein begins by letting the sides of triangle ABCbe a, b, jind c. Define k by:k ■= V{a + b + cx-a + b + cXa - b + c«a + b - c)
The radius, R, of the circumcircle is abc/k. The area
of triangle ABC is k/4. The radius, r, of the incirdeis k/2(a + b + c). Denote the side FE of triangleFED by a'. Then a' can be expressed in terms of rand angle A:
(a')J = 2^(1 + (cos A)).
The term (cos A) can be expressed in terms of a, b,and c using the law of cosines:a2 = b2 + c3 - 2bc cos A.Combining the last two expressions:a' = (-a + btc)8x V (a - b + c)(a + b -a (a + btc)8x V (a - b + c)(a + b - cybc.with similar expressions for b' and c\ The area oftriangle FED is k'/4 where k' is defined similarly tok with a' substituted for a, etc. After much algebra,the area of triangle FED is
k(-a + b + c)(a - b + c)(a + b - c)/16abc= k'/16abc(a + b + c)= (r/2R)(area ABC).
To establish his solution, Phelps Meaker lets Hequal the altitude AD of triangle ABC; O is the water of the incircle; G is the cenler of the circumcircle-and a is one-half of the apical angle. Then, withrespect to triangle AEO
r = EO = OD = (H - r) sin a ;r + r sin a = H sin a; andr = H sin a/(l + sin a).With respect to triangle ABD-AJ = AB/2 = H/2 cos a;
JANUARY 1985
AG = AJ/COS a = H/2 cos2 a;
d = 2AC = H/cos2 a.
The area of triangle ABC is
H X H Ian a = H' sin a/cos a.
With respect to triangle EDF:
EP = EO cos a; angle PEO = a; OP = EO sin a.
Then the area of triangle EDF is given by
EP x (OD + OP) = r cos a (r + r sin a)
= r2 cos a + r1 sin a cos a = r^ cos a (1 + sin a).
Substituting for r, the area of triangle EDF is given
by
H2 sin2 a cos a (1 + sin a)/(l + sin a)!= H2 sin2 a cos a/(l + sin a).
(r/d)A = (H sin a/(l + sin a)l[cos2 a/H][H2 sin a/
cos a]
= H2 sin2 a cos a/(l + sin a).Also solved by Matthew Fountain, Richard Hess,
David Evans, Robert Hollenbach, Henry Lieber-
man, Mary Linderman, Howard Stern, and the pro
poser, Harry Zaremba.
Better Late Than Never
1983 JUL 5. Matthew Fountain sent us the follow
ing:
Donald Savage's comments on 1983 JUL 5 stimu
lated me to further investigation. I found (a) there
are 29(2)" n-digit numbers whose squares are suit
able with respect to their last n digits and (b) there
are 29(2)n(0.2)' n-digit numbers whose squares are
suitable with respect to their last n + r digits; (a) is
exact when n>3; (b) is an excellent approximation
when n is large and r small. Both (a) and (b) are
consistent with the notion that the middle digits of
squares are representative of random numbers, ex
cept that (a) is more than chance. I wrote a Pascal
program that generated all the (a) numbers up to
and including those of 20 digits and printed out
those that came closest to having entirely suitable
squares. My IBM took about 2.5 days to cover the
range of 15 through 20 digits. 1 found that the results
differed from (b) because of the pecularity of two
sequences of digits. For example, (b) expects there
to be 3.1 numbers of 20 digits or less with squares
having their last 30 digits suitable. Actually there
are 22. But 15 differ from 83,333,333,333,333,333,332
only in the first six or less digits, and four differ
from 21,666,666,666,666,666,662 in the first five or
less digits. Only 78,537,356,970,849,674,736 appears
to resemble a random number. My conclusion is
that (b) is probably a good estimate of the odds that
there exists a large n-digit number with an (n + r)-
digit square, all of whose digits are suitable. The
pecularity of the two sequences should not affect
(b) when r = n or r = (n - 1). The odds, of course,
seem very small. It is interesting that Los Alamos
in its early days took sequences of digits from the
middles of consecutive squares as random num
bers, a suggestion of von Neumann. Some cyclic
patterns were observed, the worst being too many
consecutive zeroes.
19S4 APR 1. Harry (Hap) Hazard points out that
spades can be led or steel but not lead.
APR 2. Richard Halloran has responded.
M/J.5. Dayton Datlowe has responded.
JUL 2. Ivor Morgan and Mary Lindenberg have re
sponded.
A/S SD2. Arthur Carp, Ronald Martin, and Gordon
Thomas report that pan's dimensions are 7 x 5 x
1.5 inches.
Proposers' Solutions to Speed Problems
SD 1. The editor needs to assume that the speed
ratio between Joe and Bill is constant for wrapping
and addressing. In that case, the ratio is V3 and
Joe should receive
$10/(1 + V3).
SD 2. Lead low toward the A], and hope West has
the *Q. (My partner found this line of play, but
most declarers went down.)
AlexanderKusko, Inc
Rosoarch,
development and
engineering services
in the electrical
engineering field
Specialties:
Electric power
systems
Electric transportation
oqulpmont
Electric machinery
and magnetics
Solid-state motor
drives, rectifiers,
Inverters
Feedback control
systems
Computer applications
and modoling
Evaluation,
Investigation,
patents.
Alexander Kusko '44
161 Highland Avenue
Needham Heights, MA
02194
(617)444-1361
Pkxagli&Associates
Corporate
Health Coro
Consultants
Giorgio Plccagll,
Ph.D., '67
100 Dorado Terrace
San Francisco, CA
94112
(415) 333-5084
TheGodman
Company,Inc
Industrial and
Commercial Real Estate
Mark Gottesman '70
(M.C.P.)
211 Congress Street
Boston, MA 02110
(617) 423-6500
James
Goldstein&
Partners
ARCHITECTSENGINEERSPLANNERS
S. James Goldstein '46
Eliot W. Goldstein '77
225 Mlllburn Avenue
Mlllburn, NJ 07041
(201)467-6840
SERVING THE SCIENCE COMMUNITY SINCE 1953
FACILITIESResearch &
Development
Education & Training
Management &
Support
Manufacturing &
Warehousing
FOR HIGH
TECHNOLOGY
FIELDS
Biochemistry
Chemical Engineering
Chemistry
Computer Science
Electronics
EMI Structures
Hazardous Materials
Information Science
Laboratory Animals
Lasers
Medical Sciences
Microelectronics
Monoclonal
Antlbodlos
Particle Accelerators
Pulp * Paper
Recombinant ONA
Solid State Physics
Telecommunications
Toxicology
Wind Tunnels
TechnologyMarketingGroup Inc.
Marketing Consultants
for Technology
Companies
Extending Your
Resources
to Achieve:
• Sales Growth
• New Markets
• New Applications
• Product
Enhancements
• Profitable Pricing
Lesllo C. Hruby, SM 73
F. Michael Hruby
326 Burroughs Rd.
Boxborough,MA01719
(617)263-0646
ThomasKDyer,
Inc.
DIVISION OF HNTB
Consulting Engineers
Rail Transportation
Thomas K. Dyer '43
1762 Massachusetts Ave.
Lexington, MA 02173
(617) 862-207S
Washington, D.C.
(202) 466-7755
Chicago, IL
(312)663-1575
Philadelphia, PA
(215)569-1795
TECHNOLOGY REVIEW A27
Massachusetts Institute
of Technology
Report of the PresidentFor the Academic Year
1983-1984
The idea of MIT began in the mind of one man —
William Barton Rogers— and took root in a fertile setting:
in nineteenth-century America, in Boston. It was an idea
that captured the imagination and intellectual energy of
those citizens who saw the need for a new kind of ed
ucation which would emphasize, in Rogers's words,
"... the value of science in its great modern applications
to the practical arts of life, to human comfort, and health,
and to social wealth and power." The founding idea
was accompanied by an impulse, a spirit, which was
just as revolutionary as the idea itself: the spirit of in
venting the future. Rogers was not bound by the tra
ditional ways of organizing and seeking knowledge; his
entrepreneurial spirit and willingness to take risks in
pursuit of an exciting idea are the MIT tradition.
As MIT has grown, it has embraced and invented a
host of intellectual domains. Our departments and
academic programs reflect in their names and in their
activities the remarkable ability of the faculty to anticipate
and shape the future. Scores of interdepartmental re
search laboratories and centers constitute perhaps an
even more sensitive barometer of the ways in which
MIT transcends the boundaries of tradition in both or
ganization and style.
Today, our fields of study and scholarship include
management, urban planning, humanities and the social
sciences, as well as the natural sciences, architecture,
and engineering. MIT has become more than an institute
of technology in the nineteenth-century sense of that
phrase. To use the words of James R. Killian, Jr., first
spoken in 1949, MIT is "a university polarized around
science, engineering, and the arts." And yet the founding
idea of MIT continues and is central to our future, for
we are now and must remain the strongest science-based
research university in the world. Our historic commit
ment to scientific and quantitative methods remains at
the core of our approach to learning; it permeates the
whole spectrum of our degree programs, and is a touch
stone of common interest and purpose for all who study,
teach, and work here.
Given this foundation, MIT's current mission can be
stated succinctly: to provide the highest quality programs
of education and research in each of those areas of study
and investigation in which we have developed strength
and competence, and to do so with a strong commitment
to public service and to a diversity of backgrounds, in
terests, and points of view among the faculty, students,
and staff.
A28 JANUARY 1985 PHOTOGRAPH OF WILLIAM BARTON ROGERS: M.IT MUSEUM